Strengthening of Reinforced Concrete Structures: Using Externally-Bonded FRP Composites in...

Contents iii

Strengtheningof reinforced

concretestructures

Using externally-bonded FRPcomposites in structural and

civil engineeringEdited by L C Hollaway and M B Leeming

CRC PressBoca Raton Boston New York Washington, DC

Cambridge England

SORPR 12/19/98, 2:33 PM3

iv Contents

Published by Woodhead Publishing Limited, Abington Hall, AbingtonCambridge CB1 6AH, England

Published in North and South America by CRC Press LLC,2000 Corporate Blvd, NWBoca Raton FL 33431, USA

First published 1999, Woodhead Publishing Ltd and CRC Press LLC

© 1999, Woodhead Publishing LtdThe authors have asserted their moral rights.

This book contains information obtained from authentic and highly regardedsources. Reprinted material is quoted with permission, and sources are indicated.Reasonable efforts have been made to publish reliable data and information, butthe authors and the publishers cannot assume responsibility for the validity of allmaterials. Neither the authors nor the publishers, nor anyone else associated withthis publication, shall be liable for any loss, damage or liability directly orindirectly caused or alleged to be caused by this book.

Neither this book nor any part may be reproduced or transmitted in any formor by any means, electronic or mechanical, including photocopying, microfilmingand recording, or by any information storage or retrieval system, without permis-sion in writing from the publishers.

The consent of Woodhead Publishing and CRC Press does not extend tocopying for general distribution, for promotion, for creating new works, or forresale. Specific permission must be obtained in writing from Woodhead Publish-ing or CRC Press for such copying.

Trademark notice: Product or corporate names may be trademarks or registeredtrademarks, and are used only for identification and explanation, without intentto infringe.

British Library Cataloguing in Publication DataA catalogue record for this book is available from the British Library.

Library of Congress Cataloging in Publication DataA catalog record for this book is available from the Library of Congress.

Woodhead Publishing ISBN 1 85573 378 1CRC Press ISBN 0-8493-1715-0CRC Press order number: WP1715

Cover design by The ColourStudioTypeset by Best-set Typesetter Ltd., Hong KongPrinted by TJI Digital, Padstow, Cornwall

SORPR 12/19/98, 2:33 PM4

Reprinted 2001

Contents v

Contents

Preface ixList of contributors xiii

1 Role of bonded fibre-reinforced composites instrengthening of structures 1J J DARBY

1.1 Introduction 11.2 What is ‘strengthening with bonded fibre reinforced

polymer composite plates’? 11.3 The market for strengthening 21.4 Strengthening techniques 31.5 Advantages and disadvantages of FRP composite

plate bonding 41.6 Client concerns when introducing new techniques 81.7 Risk to clients when adopting FRP composite plate

bonding 81.8 Conclusions 10

2 Review of materials and techniques for plate bonding 11L C HOLLAWAY AND M B LEEMING

2.1 Introduction 112.2 Structural adhesive bonding 112.3 External strengthening using steel plates 172.4 External strengthening using composite materials 212.5 Strengthening of reinforced concrete members in shear 342.6 Applications of FRP strengthening 362.7 Summary and conclusions of literature review 382.8 References 39

3 Materials 46A R HUTCHINSON AND J QUINN

v

SORPR 12/19/98, 2:33 PM5

vi Contents

3.1 Adhesive bonded connections 463.2 Composite materials 473.3 Adhesive materials 573.4 Adhesion and surface preparation 683.5 The bonding operation 753.6 Durability and fire 803.7 Painting 803.8 Summary 803.9 References 81

4 Structural strengthening of concrete beams usingunstressed composite plates 83L C HOLLAWAY AND G C MAYS

4.1 Introduction 83Part A Laboratory tests 84

4.2 General form and behaviour of loaded beams 844.3 Geometric parameters 914.4 Discussion 107

Part B Field investigation 1094.5 Testing programme for 18m beam 1094.6 Observations 1314.7 Concluding remarks 1324.8 References 132

5 Structural strengthening of concrete beams usingprestressed plates 135H N GARDEN AND G C MAYS

5.1 Introduction 1355.2 Review of previous prestressing studies using composite

plates 1375.3 Prestressing technique employed in the laboratory 1385.4 Results of laboratory tests for concrete beams

strengthened with prestressed plates in the ROBUSTprogramme 140

5.5 Results of field investigations of concrete beamsstrengthened with prestressed plates in the ROBUSTprogramme 146

5.6 Observations 1495.7 Concluding remarks 1545.8 References 154

6 Environmental durability 156A R HUTCHINSON AND L C HOLLAWAY

SORPR 12/19/98, 2:33 PM6

Contents vii

6.1 Introduction 1566.2 Environmental and service conditions 1576.3 Factors affecting joint durability 1586.4 Environmental durability of adhesive bonded joints 1606.5 Procedures for assessing environmental effects

on materials and on bonded joints 1636.6 Effect of environment on the component materials used

in the ROBUST system 1666.7 Influence of surface treatment and effects of environment

on joints and interfaces 1736.8 Other factors affecting service performance 1796.9 Summary 1816.10 References 181

7 Time-dependent behaviour and fatigue 183R A BARNES AND H N GARDEN

7.1 Introduction 183PART A Time-dependent behaviour 183

7.2 Introduction 1837.3 Time-dependent characteristics of concrete 1847.4 Time-dependent characteristics of steel 1847.5 Time-dependent characteristics of adhesives 1847.6 Time-dependent characteristics of plated beams using

steel plates 1927.7 Time-dependent characteristics of FRP component

materials and FRP composites 1947.8 Time-dependent characteristics of plated beams using

polymer composite plates 1967.9 Creep tests conducted during the ROBUST project 197

PART B Fatigue behaviour 2007.10 Introduction 2007.11 Fatigue of unplated beams 2017.12 Fatigue of adhesives 2037.13 Fatigue of FRP materials 2067.14 Fatigue of plated beams using steel plates 2087.15 Fatigue of short span plated beams using FRP plates 2117.16 Fatigue of long span plated 2.3 m beams using FRP plates 2137.17 Concluding summary 2177.18 References 218

8 Analytical and numerical solutions to structuralstrengthening of beams by plate bonding 222P S LUKE

SORPR 12/19/98, 2:33 PM7

viii Contents

8.1 Introduction 2228.2 Classical analysis 2238.3 Finite element analysis 2238.4 Effect of adhesive material 2318.5 Prestressed 18.0m concrete beams 2318.6 Beams with unstressed plates 2348.7 Beams with stressed plates 2378.8 Concluding remarks 2408.9 References 2418.10 Bibliography 241

9 Design and specifications for FRP plate bondingof beams 242M B LEEMING AND J J DARBY

9.1 Introduction 2429.2 Practical design rules and guidelines 2429.3 Application of the technique 2579.4 Materials 2589.5 Workmanship 2609.6 Quality control 2619.7 In-service inspection and maintenance 2669.8 References 2679.9 Bibliography 268

10 Site construction techniques 270A P R IMOLDI

10.1 Introduction 27010.2 Steel plate bonding 27010.3 Adhesive bonding of carbon fibre composite plates –

site requirements 27210.4 Economics 28610.5 Conclusion 28710.6 References 287

11 Case studies of carbon fibre bonding worldwide 288M A SHAW AND J F DREWETT

11.1 Introduction 28811.2 System properties 28911.3 Case histories 29011.4 References 324

Index 325

SORPR 12/19/98, 2:33 PM8

Role of bonded fibre-reinforced composites 1

1

1Role of bonded fibre-reinforced composites

in strengthening of structures

J J D A R B Y

1.1 Introduction

The authors contributing to this volume have been immersed in the devel-opment of advanced composite materials for strengthening structures for anumber of years. Yet, in 1998, this can still be described as a new technique,with the total number of applications worldwide measured at most in hun-dreds. From this cautious beginning, the author believes that a rapid expan-sion in usage will take place as the benefits are more widely realised. Allclients and designers seek solutions that are durable and cost effective,exactly those requirements which fibre reinforced composite strengtheningsystems can be designed to meet. However, clients must also gain trust innew techniques before they will be willing to adopt them. That trust must befirmly based on an understanding of material behaviour, the design processand the risks of implementation. It is hoped that this book will assist in thatprocess, particularly by disseminating some of the knowledge that has beengained during the ‘ROBUST’ research project. Inevitably it will take timeto foster a wide appreciation of these new materials amongst the construc-tion community. It will not be assumed that readers have any previousexperience of composite materials. All aspects of composite plate bondingare covered in some detail in individual chapters, but a more generalintroduction to the techniques is appropriate first. This will take the form ofa definition of terms.

1.2 What is ‘strengthening with bonded fibre

reinforced polymer composite plates’?

• Fibre reinforced polymer (FRP) composites: FRP composites comprisefibres of high tensile strength within a polymer matrix. The fibresare generally carbon or glass, in a matrix such as vinylester orepoxy. These materials are preformed to form plates under factoryconditions, generally by the pultrusion process. For experimentation,

SOR1 12/19/98, 2:34 PM1

2 Strengthening of reinforced concrete structures

plates may be manufactured in smaller quantities from pre-impregnatedfibre mats.

• Bonded plates: the preformed plates are fixed externally to the structurewith adhesives, usually of epoxy, to promote composite structural ac-tion, although additional mechanical fixings may be used if deemednecessary by particular circumstances.

• Structural strengthening: the load bearing capacity of structures maybe increased or restored, either locally or overall. Plates may be in-stalled unstressed, or stressed on site effectively to prestress the struc-ture. Most experimental work has been undertaken by applyingcomposites to concrete, but timber, stone, steel or cast iron may also bestrengthened.

1.3 The market for strengthening

Modern civilisation relies upon the continuing performance of a widevariety of structures, ranging from industrial buildings and power stationsto bridges. Although these structures may appear very different, theirmanagers are likely to recognise a number of common features:

• structural deterioration perhaps increased by environmental factors• changes in use or imposed loading• the need to minimise closure or disruption during repairs• the need to extend useful life whilst minimising capital outlay• more stringent financial disciplines requiring the evaluation of the

whole life cost of solutions.

The number of structures in the world continues to increase, as does theiraverage age. The need for increased maintenance is inevitable. Completereplacement is likely to become an increasing financial burden and is cer-tainly a waste of natural resources if upgrading is a viable alternative. Theway in which FRP composite plate bonding can help will be illustrated byconsidering two particular structure types, buildings and bridges.

• Buildings: industrial buildings may be adapted for new uses, increasingfloor or slab loading. Externally bonded plates will increase capacitywith negligible increase in construction depth. Structural alterationsmay require removal of columns or holes to be cut through slabsfor purposes such as new lifts or services. External reinforcementin these circumstances may be the only alternative to partial demoli-tion and replacement, with all the disruption to production which thatentails.

• Bridges: loads on bridges are increasing, due to increases in the permit-ted vehicle weights as well as the volume of traffic. At the same time

SOR1 12/19/98, 2:34 PM2


material deterioration is becoming more evident, particularly that dueto reinforcement corrosion induced by contamination with de-icingsalts. For this reason, a large scale assessment programme is underwayin the UK to examine the load capacity of all bridges of uncertainstrength. This has already revealed the need for extensive strengthen-ing. FRP composite plate bonding will offer the best solution for manyof these structures, particularly where short construction periods may bea key factor.

Cost is probably the most influential factor when assessing the merits ofalternative methods. Detailed costing would be out of place in a book ofthis kind and would date quickly. This is particularly the case for newtechniques, as prices can be expected to fall as more material suppliers andcontractors enter the growing market. However, the case for bonded fibrereinforced composites can best be illustrated by the fact that these materialsare already winning competitive tenders against alternative solutions.

1.4 Strengthening techniques

The art of designing strengthening schemes with FRP composite platebonding is at an early stage. Detailed guidance on what reinforcement anddetailing should be used in every particular circumstance cannot be pro-vided. Economical solutions depend upon an understanding of the materi-als and experience of what they can safely achieve. There are many optionsopen to the skilful designer. Just as reinforced concrete may be designed tobehave differently according to the mix of concrete and reinforcement, sothe composite plates may use different reinforcement materials, in differentproportions, and within different matrix materials. These plates may thenbe of different lengths, and multiple layers may be used. These may be fixedat any required geometry on the surface of the structure. The adhesives andsurface preparation may vary. The plates may be stressed or unstressed andthe ends mechanically anchored or bonded by adhesive only.

This wide range of options must be seen as an advantage and as anopportunity for the knowledgeable designer to tailor the strengtheningscheme to the needs of the particular structure. But there is also a potentialdanger arising from application by designers without experience. It is diffi-cult at the present stage of composite plate bonding to write a specificationthat covers all potential situations. Reinforced concrete specifications arestill developing a century after the initial application of the material wasconcentrated in the hands of a number of specialist exponents. Fortunately,development of FRP composite plate bonding will be much faster. Analysismethods are available to speed up the process of understanding structuralbehaviour and we can build upon previous knowledge.

SOR1 12/19/98, 2:34 PM3


Projects such as ROBUST have provided a solid basis for designers to useFRP composite plate bonding to enhance flexural behaviour, using bothstressed and unstressed plates. Much has also been learned about the needfor fixing of the plates due to end peel effects. Shear strengthening, on theother hand, has been little researched to date.

The range of structural needs and deficiencies for which FRP compositeplate bonding already offers an appropriate solution is very wide, as illus-trated in Table 1.1.

1.5 Advantages and disadvantages of FRP composite

plate bonding

All structural problems have more than one technical solution, and finalselection will ultimately rest upon an economic evaluation of the alterna-tives. Enlightened clients will ensure that this evaluation includes an esti-mate of the total costs that will be incurred during the required service life,rather than selection of the scheme with the minimum initial cost. The totalcosts will include future maintenance, as well as all consequential costs suchas loss of production or traffic delay costs.

The most obvious technical solution with which to compare FRP compos-ite plate bonding is steel plate bonding, as many of the aspects are commonto both. Such a comparison is made below. However, FRP composite platebonding should not be thought of as simply an improved form of steelplate bonding. The new material offers such versatility that new solutionswill become practicable, particularly those arising from prestressing of theplates

The potential advantages of FRP composite plate bonding are as follows:

• Strength of plates: FRP composite plates may be designed with compo-nents to meet a particular purpose and may comprise varying propor-tions of different fibres. The ultimate strength of the plates can thus bevaried, but for strengthening schemes the ultimate strength of the platesis likely to be at least three times the ultimate strength of steel for thesame cross-sectional area.

• Weight of plates: the density of FRP composite plates is only 20% of thedensity of steel. Thus composite plates may be less than 10% of theweight of steel of the same ultimate strength. Apart from transportcosts, the biggest saving arising from this is during installation. Compos-ite plates do not require extensive jacking and support systems to moveand hold in place. The adhesives alone will support the plate until curinghas taken place. In contrast, fixing of steel plates constitutes a significantproportion of the works costs.

SOR1 12/19/98, 2:34 PM4


Table 1.1 Applications for composite plate bonding

Structural FRP composite plate Commentsneed/deficiency bonding solution

Corrosion of Replacement of lost Damaged concretereinforcement in reinforcement by plates of must be replacedreinforced concrete equivalent effect without impairing

behaviour of plates

Inadequate flexural Design FRP composite Extent of strengtheningcapacity of reinforced plate bonding solution to limited by capacity ofconcrete add tensile elements concrete in

compression. Platesanchored by bond ormechanically at theirends

Lost prestress due to Replace prestress that has Need to ensure nocorrosion in prestressed been lost with stressed overstress of concreteconcrete composites in the short term

Safety net to cover Add plates, either stressed Method may beuncertain durability of or unstressed, to ensure particularly appropriateprestressed concrete safety. Particularly with segmental

appropriate if corrosion construction. May beunlikely but possible combined with a

monitoring system

Inadequate stiffness Add external prestress byor serviceability of means of a stressedcracked reinforced composite plateconcrete structure

Potential overstress Analyse stresses due todue to required alteration, and designstructural alteration composite reinforcement

before removing load-bearing members

Avoidance of sudden Addition of FRP compositefailure by cracking of plate bonding, eithercast iron stressed or unstressed, to

tensile face

Increase in structural Increase in stiffness and Particularly appropriatecapacity of timber ultimate capacity by plate with historic structuresstructures bonding

Enhancement of Enhanced by external Web reinforcementshear capacity bonding of stressed plates, techniques little

or by web reinforcement researched

SOR1 12/19/98, 2:34 PM5


• Transport of plates: the weight of plates is so low that a 20m longcomposite plate may be carried on site by a single man. Some platesmay also be bent into a coil as small as 1.5 m diameter, and thus may betransported in a car or van without the need for lorries or subsequentcraneage facilities. The flexibility of plates enables strengtheningschemes to be completed within confined spaces.

• Versatile design of systems: steel plates are limited in length by theirweight and handling difficulties. Welding in situ is not possible, becauseof damage to adhesives, and expensive fixing of lap plates is thereforerequired. In contrast, composite plates are of unlimited length, may befixed in layers to suit strengthening requirements, and are so thin thatfixing in two directions may be accommodated by varying the adhesivethickness.

• Easy and reliable surface preparation: steel plates require preparation bygrit blasting, followed by careful protection until shortly beforeinstallation. In contrast, the ROBUST project has demonstrated thatcomposite plates may be produced with a peel-ply protective layer thatmay be easily stripped off just before the adhesive is applied.

• Reduced mechanical fixing: composite plates are much thinner than steelplates of equivalent capacity. This reduces peeling effects at the ends ofthe plates and thus reduces the likelihood of a need for end fixing. Theoverall depth of the strengthening scheme is reduced, increasing head-room and improving appearance.

• Durability of strengthening system: there is the possibility of corrosionon the bonded face of steel plates, particularly if the concrete to whichthey are fixed is cracked or chloride contaminated. This could reducethe long term bond. Composite plates do not suffer from such deterio-ration.

• Improved fire resistance: composite plates are a low conductor of heatwhen compared with steel, thus reducing the effect fire has on theunderlying adhesives. The composite itself chars rather than burns andthe system thus remains effective for a much longer period than steelplate bonding.

• Reduced risk of freeze/thaw damage: there is theoretical risk of waterbecoming trapped behind plate systems, although this should not occurif they are properly installed. In practice, this has not been found tobe a problem. However, if water did become trapped in this way, theinsulating properties of the composite materials would reduce the risk ofdisruption of the concrete due to freeze/thaw. Loss of bond would alsobe evident by tapping the composite, but would be more difficult todetect with steel.

• Maintenance of strengthening system: steel plates will require mainte-nance painting and may incur traffic disruption and access costs as well

SOR1 12/19/98, 2:34 PM6


as the works costs. Composite plates will not require such maintenance,reducing the whole life cost of this system.

• Reduced construction period: many of the practical advantages de-scribed above combine to enable composite plates to be installed ingreatly reduced time periods when compared with steel plates. As wellas lower contract costs, the traffic delay costs are minimised. Installationfrom mobile platforms becomes possible and it may become practicableto confine work within such restraints as limited railway possessions ornight-time working.

• Ability to prestress: the ability to prestress composites opens up awhole new range of applications for plate bonding. The plate bondingmay be used to replace lost prestress and the shear capacity of sectionswill be increased by the longitudinal stresses induced. Formationof cracks will be inhibited and the serviceability of the structure en-hanced. Strengthening of materials such as cast iron also becomes morepracticable.

The potential disadvantages of FRP composite plate bonding are asfollows:

• Cost of plates: fibre reinforced composite plates are more ex-pensive than steel plates of the equivalent load capacity. However, thedifference between the two materials is likely to be reduced as pro-duction volumes and competition between manufacturers increases.Comparison of total contract costs for alternative methods of strength-ening will be based on labour and access costs as well as materialcosts. Open competition has already shown that FRP compositeplate bonding is the most economic solution in virtually all testedcases, without taking into account additional advantages such asdurability.

• Mechanical damage: FRP composite plates are more susceptible todamage than steel plates and could be damaged by a determined attack,such as with an axe. In vulnerable areas with public access, the risk maybe removed by covering the plate bonding with a render coat. Fortu-nately, if damage should occur to exposed FRP composite plate, such asby a high load, repairs can be undertaken much more easily than with asteel plate. A steel plate may be dislodged, or bond broken over a largearea, which would damage bolt fixings and necessitate complete re-moval and replacement. However, with FRP composite plate bondingthe damage is more likely to be localised, as the plate is thinner andmore flexible. With FRP composite, the plate may be cut out over thedamaged length, and a new plate bonded over the top with an appropri-ate lap.

SOR1 12/19/98, 2:34 PM7


1.6 Client concerns when introducing new

techniques

The construction industry is cautious over the introduction of any newtechnique, and this is understandable in view of problems with earliermaterials. Even concrete is now known to deteriorate in many more waysthan was widely appreciated just 30 years ago, and clients have becomemuch more aware of the durability and maintenance implications of thesolutions they adopt.

New techniques inevitably involve risk, because there cannot be a longtrack record of successful performance. Reassurance must be gained inother ways. Risk can still be managed and evaluated, a process which hasbeen made much easier by advanced technology. We have the benefit ofconsiderable understanding of design parameters and deterioration mecha-nisms. Finite element tools for analysing structural behaviour are morepowerful than ever before. Thus we have the opportunity to analyse poten-tial problems in advance and then to extend applications only within theboundaries of understanding.

The ROBUST project demonstrates what can be achieved. At a cost ofabout £1 million, a large number of strength, fatigue and durability testshave been performed. The expertise of designers, researchers, contractorsand clients has been combined to investigate those risks that may resultfrom applying the new technology. The volume of research work on com-posite plate bonding exceeds that which was undertaken before steel platebonding was introduced and which was deemed sufficient to accept steelplate bonding in public works. Significantly, the replacement of steel bycomposite materials is a smaller technological change than was the first useof external reinforcement. The extensive research work on composites wasfully justified and reflects the fact that clients in the present climate needreassurance, particularly where structural safety is involved.

1.7 Risk to clients when adopting FRP composite

plate bonding

The decision by clients to use FRP composite plate bonding will rest uponbalancing the clear advantages arising from use of the material, which havebeen indicated above, against the risk of potential problems. Those riskshave been minimised by the work to date, and could be summarised asfollows:

• Durability of carbon fibre reinforcement: carbon is a basic element oc-curring naturally in the environment. There is no known degradationmechanism.

SOR1 12/19/98, 2:34 PM8


• Durability of glass fibre reinforcement: E-glass is attacked by alkalinityfrom concrete if the fibre comes into direct contact with it, although thismay be overcome to some extent by use of Z-glass; it should be men-tioned, however, that the laminating polymer protects the fibre fromdirect contact with the concrete in the plate bonding technique. Stresscorrosion may also be a problem with glass under continuous high statesof stress. For these reasons, and also because the ultimate strength ofglass is less than carbon, it is not advisable to use this material for mainreinforcement without further research, despite its lower cost.

• Durability of the composite matrix: vinylester has been used in thechemical process industry for several decades as the preferred corrosionresistance barrier against a multiplicity of highly corrosive chemicals.Sunlight can cause some yellowing and surface degradation if no ultra-violet stabilisers are introduced into the formulation, but tensile andflexural strengths and moduli are not found to be significantly affected.

• FRP composite plate performance: carbon fibre reinforced compositeshave been used for 20 years in highly stressed areas in commercial andmilitary aircraft and racing cars. The construction industry makes muchlower demands on the performance of the material.

• Adhesive performance: many composite aircraft wings are glued to themain fuselage, which demonstrates the structural performance of com-posites and compatible adhesives. The adhesives used for compositeplate bonding have been used in the construction industry for 20 yearsand have no known degradation mechanism. Performance with steelplate bonding has been accepted, and materials testing has shown theadhesives to be equally effective with composite materials. Contamina-tion must be avoided, but peel ply reduces risk of poor workmanship.Moisture within the concrete has not been found to be a problem.

• System performance: the fatigue properties of FRP composite platebonding have been found to be excellent, with fatigue failure not beinginitiated by any plate bonding component. Instead, tests have shown thefatigue life of reinforced concrete beams to be limited by the fatiguelife of the embedded reinforcement. The addition of plate bonding willtherefore only be significant if it increases the stress range of the rein-forcement. Creep properties of carbon fibre are excellent and creep ofthe adhesive has not been found to be a problem.

• System design: much has been learned about the failure modes of beamsreinforced by composite plate bonding and computer predictions arenow remarkably accurate. Nevertheless, this is clearly a complicatedarea. Applications will inevitably extend to structures with features thatdiffer in certain aspects from those tested. A period of monitored appli-cation is required before an all-embracing specification can be pro-duced. The risks in the interim are minimal if design is undertaken by

SOR1 12/19/98, 2:34 PM9


those with an understanding of the behaviour of composite materials,who are thus able to recognise the need for caution and to invokeconservative features when the situation demands. The risk is higher ifFRP composite plate bonding is designed by those without sufficientbackground knowledge, as deficiencies will not necessarily show up atthe time of construction.

1.8 Conclusions

This introductory chapter can do no more than summarise the presentposition. It is within the later chapters that the progress made in researchingFRP composite plate bonding can be illustrated in detail. Clients and de-signers will want to know whether the time is now right for them to utilisecomposite plate bonding techniques. There will undoubtedly be variation inthe willingness to adopt new solutions, but it is to be hoped that the textwithin this volume will support the following conclusions:

• Fibre reinforced composite plate bonding offers significant advantagesover steel plate bonding for the vast majority of strengthening applica-tions.

• Fibre reinforced composite plate bonding is so versatile that the rangeof applications for which external reinforcing is appropriate will in-crease significantly.

• No construction or repair method involving structural analysis and dete-rioration mechanisms can be said to be completely understood, includ-ing all of those currently in everyday use. However, FRP compositeplate bonding has been sufficiently researched to enable the techniquesto be applied confidently on site, providing care is taken.

• The method of FRP composite plate bonding is here to stay and isalready being actively marketed. The number of applications worldwideis set to grow very fast. The challenge is to ensure that these applicationstake full account of the current state of knowledge. The benefits mustnot be put at risk by inappropriate or badly detailed applications under-taken by the inexperienced.

SOR1 12/19/98, 2:34 PM10

Review of materials and techniques for plate bonding 11

11

2Review of materials and techniques for

plate bonding

L C H O L L A W A Y A N D M B L E E M I N G

2.1 Introduction

This chapter provides an introduction to the flexural rehabilitation orstrengthening of reinforced concrete (RC), prestressed concrete and steelmembers using externally bonded steel or fibre reinforced polymer (FRP)composites plates by reviewing the most significant investigations reportedin the literature. In addition, a section is devoted to the strengthening of RCmembers in shear utilising FRP plates. However, since the external platingand its application as a strengthening technique has only been made possi-ble by the development of suitable adhesives, consideration is initially givento the types of adhesive which may be used for external plate bonding andtheir requirements for this application. After considering reported platebonding studies, a brief review of surface preparation techniques applicableto FRP and concrete adherends is presented.

2.2 Structural adhesive bonding

Structural adhesives are generally accepted to be monomer compositeswhich polymerise to give fairly stiff and strong adhesive uniting relativelyrigid adherends to form a load-bearing joint (Shields, 1985). The feasibilityof bonding concrete with epoxy resins was first demonstrated in the late1940s (ACI, 1973), and the early development of structural adhesives isrecorded by Fleming and King (1967). Since the early 1950s adhesives havebecome widely used in civil engineering (Mays, 1985). However, althoughthe building and construction industries represent some of the largest usersof adhesive materials, many applications are non-structural in the sense thatthe bonded assemblies are not used to transmit or sustain significantstresses (e.g. crack injection and sealing, skid-resistant layers, bonding newconcrete to old). Truly structural application implies that the adhesive isused to provide a shear connection between similar or dissimilar materials,enabling the components being bonded to act as a composite structural unit.A comprehensive review of applications involving the use of adhesives in

SOR2 12/19/98, 2:39 PM11


civil engineering is given by Hewlett and Shaw (1977), Tabor (1982) andMays and Hutchinson (1992).

Assessment of an adhesive as a suitable product for structural use musttake into account the design spectrum of loads, the strength and stiffnessof the material under short term, sustained or cyclic loads and the effecton these properties of temperature, moisture and other environmentalconditions during service (Mays, 1993). Concern regarding the durabilityproperties of adhesive joints has meant that resistance to creep, fatigueand fracture are considered of greater importance than particularly highstrength (Vardy and Hutchinson, 1986). Temperature is important at allstages in the use and performance of adhesives, affecting viscosity andtherefore workability, usable life and contact time, rate of cure, degree ofcross-linking and final cured performance (Tu and Kruger, 1996). Control-led conditions are therefore generally required during bonding. This appliesequally during the surface treatment procedures if a durable system is to beachieved. Adhesives, which are workable and cure at ambient tempera-tures, have been used and are able to tolerate a certain amount of moisturewithout a marked reduction in performance. These must have adequateusable time under site conditions and a cure rate which does not hinder theconstruction programme. Workmanship under conditions prevalent on siteis less conducive to quality control than in other industries, and thus abilityto tolerate minor variations in proportioning and mixing, as well as imper-fect surface treatment, is important. In addition, the products involvedare more toxic, require more careful storage and, bulk for bulk, are con-siderably more expensive than traditional construction materials. Non-destructive test methods for assessing the integrity of bonded joints are nowavailable for civil engineering applications.

Despite some drawbacks, structural adhesives have enormous potentialin future construction applications, particularly where the combination ofthick bondlines, ambient temperature curing and the need to unite dissimi-lar materials with a relatively high strength joint are important (Mays andHutchinson, 1992).

2.2.1 Type of structural adhesives

The principal structural adhesives specifically formulated for use in theconstruction industry are epoxy and unsaturated polyester resin systems,both thermosetting polymers. The formulation of adhesives is considered indetail by Wake (1982), whilst Tabor (1978) offers guidance on the effectiveuse of epoxy and polyester resins for civil engineering structures.

Two-part epoxies, first developed in the 1940s (Lee and Neville, 1967),consist of a resin, a hardener or cross-linking agent which causes polymeri-sation, and various additives such as fillers, tougheners or flexibilisers, all of

SOR2 12/19/98, 2:39 PM12


which contribute to the physical and mechanical properties of the resultingadhesive. Formulations can be varied to allow curing at ambient tempera-ture, the so-called cold cure epoxies, the most common hardeners for whichare aliphatic polyamines, whose use results in hardened adhesives which arerigid and provide good resistance to chemicals, solvents and water (Maysand Hutchinson, 1992). Correct proportioning and thorough mixing areimperative when using epoxy resin systems. The rate of curing doubles asthe temperature increases by 10 °C and halves as the temperature drops by10 °C and many of the formulations stop curing altogether below a tempera-ture of 5 °C. Fillers, generally inert materials such as sand or silica, may beused to reduce cost, creep and shrinkage, reduce exotherm and the coeffi-cient of thermal expansion, and assist corrosion inhibition and fire retarda-tion. Fillers increase the viscosity of the freshly mixed system but impartthixotropy, which is useful in application to vertical surfaces.

Unmodified epoxy systems tend to be brittle when cleavage or peel forcesare imposed. Toughening of the cured adhesive can be achieved by theinclusion of a dispersed rubbery phase which absorbs energy and preventscrack propagation. Epoxies are generally tolerant of many surface andenvironmental conditions and possess relatively high strength. They arepreferred for bonding to concrete since, of all adhesives, they have a par-ticularly high tolerance of the alkalinity of concrete, as well as moisture. Bysuitable formulation, their ability to wet out the substrate surfaces can evenbe achieved in the presence of water, the resin being able to disperse thewater from the surface being bonded (Tabor, 1978).

Unsaturated polyester resins were discovered in the mid-1930s and haveadhesive properties obtained by cross-linking using a curing agent. They arechemically much more simple than epoxy resins but have a 10% contractionby volume during curing due to a volume change during the transition fromthe uncured liquid phase to the hardened resin resulting in further curingshrinkage. As a result of these factors, there are usually strict limits on thevolume of material that can be mixed and applied at any one time and as ageneral rule polyester resins do not form as strong adhesive bonds as doepoxy resins. In storage, the polyester resins are also somewhat less stableand present a greater fire hazard than epoxies. These limitations signifi-cantly restrict their applications.

The advantages of epoxy resins over other polymers as adhesive agentsfor civil engineering use can be summarised as follows (Mays andHutchinson, 1992):

• High surface activity and good wetting properties for a variety ofsubstrates.

• May be formulated to have a long open time (the time between mixingand closing of the joint).

SOR2 12/19/98, 2:39 PM13


• High cured cohesive strength, so the joint failure may be dictated by theadherend strength, particularly with concrete substrates.

• May be toughened by the inclusion of a dispersed rubbery phase.• Minimal shrinkage on curing, reducing bondline strain and allowing the

bonding of large areas with only contact pressure.• Low creep and superior strength retention under sustained load.• Can be thixotropic for application to vertical surfaces.• Able to accommodate irregular or thick bondlines.• Formulation can be readily modified by blending with a variety of

materials to achieve desirable properties.

These various modifications make epoxy adhesives relatively expensivein comparison to other adhesives. However, the toughness, range of visco-sity and curing conditions, good handling characteristics, high adhesivestrength, inertness, low shrinkage and resistance to chemicals have meantthat epoxy adhesives have found many applications in construction, forexample, repair materials, coatings and as structural and non-structuraladhesives.

2.2.2 Requirements of the adhesive for plate bonding

There are many features of an adhesive product, in addition to its purelyadhesive properties, which will form the basis for the selection of a particu-lar bonding system. Mays (1985) has considered requirements for adhesivesto be used for external plate bonding to bridges under conditions preva-lent in the UK. These requirements are extended and refined in a laterpublication referred to as a proposed Compliance Spectrum (Mays andHutchinson, 1988), which addresses the general engineering requirementsof adhesives, bonding procedures and test methods for structural steel-to-concrete bonding, based on research work at the University of Dundee(Hutchinson, 1986). The requirements proposed for the adhesive itselfcan be considered to be equally applicable to fibre reinforced polymer(FRP) plate bonding. An epoxy resin and polyamine hardener arerecommended.

Choice of a suitable adhesive is only one of a number of requirements fora successfully bonded joint. Other factors also affect the joint strength andperformance (Mays and Hutchinson, 1988) namely:

• appropriate design of the joint• adequate preparation of the adherend surfaces• controlled fabrication of the joint• protection from unacceptably hostile conditions in service• postbonding quality assurance.

SOR2 12/19/98, 2:39 PM14


Both short term and long term structural performance are likely to beimproved by using an appropriately designed joint and suitably preparingthe surface of the substrate materials. A review of factors important to thesatisfactory design of joints is given by Adams and Wake (1984) and Lees(1985) and will not be considered here. Full account must be taken of thepoor resistance of adhesives to peel and cleavage forces; shear strengthitself is unlikely to be a limiting factor. With concrete structures, the tensile/shear, or tear-off strength of the concrete should be the critical design factorif a suitable adhesive formulation is selected and appropriate methods ofsurface preparation implemented. This has been demonstrated throughdetailed shear testing on site and in the laboratory (Moustafa, 1974;Hugenschmidt, 1975; Schultz, 1976).

2.2.3 Tests to measure structural adhesive bond strength

A number of tests are available for testing adhesive and thin films (Adamsand Wake, 1984; Kinloch, 1987). However, appropriate tests for assessingbond strength in construction are complicated by the fact that the loadingcondition in service is difficult to simulate, and one of the adherends,namely concrete, tends to be weaker in tension and shear than the adhe-sives which may be used, making discrimination between adhesive systemsdifficult. As a result, confirmation of the suitability of a proposed adhesivesystem is generally limited to demonstrating that, when the bondline isstressed in the test configuration chosen, the failure surface occurs withinthe concrete substrate. Such tests may also be used to exhibit the adequacyof the surface preparation techniques employed, since it is difficult to sepa-rate the individual effects on adhesion of the adhesive type and method ofsurface treatment.

Several possible test methods have evolved to measure the bond strengthbetween adhesive and concrete substrates, mainly for applications in con-crete repair (Franke, 1986; Naderi et al., 1986). The Réunion Internationaledes Laboratoires d’Essais et de Recherches sur les Matériaux et les Con-structions (RILEM) Technical Committee 52-RAC lists some currentlyused laboratory and field test methods for assessing the bond between resinand concrete (Sasse and Friebrich, 1983). Procedures are mentioned on thestrength of adhesion in tension, shear and bending, as well as shrinkage andthermal compatibility in the context of coatings, concrete repair, concrete/concrete and steel/concrete bonds.

Variations of the slant shear test (Kreigh, 1976), in which two portions ofa standard cylinder or prism are joined by a diagonal bondline and thentested in compression, have been found to produce discriminating andconsistent results (Kreigh, 1976; Naderi, 1985; Wall et al., 1986). Tu andKruger (1996) used such a configuration to demonstrate that a flexible,

SOR2 12/19/98, 2:39 PM15


tough epoxy provided improved adhesion compared to a more brittle mate-rial because it allows redistribution of forces before fracture. However,Tabor (1985) concluded that the slant shear test is of little use in assessingadhesion between resin and concrete because the interfaces are not sub-jected to tensile forces.

In assessing the shear connection in steel/concrete composite construc-tion, tests at the Wolfson Bridge Research Unit at the University of Dundeeemployed a kind of double-lap joint configuration as described by Solomon(1976), in which fracture was characterised by shear failure of the concreteadjacent to the interface with the adhesive.

The University of Surrey (Quantrill et al., 1995) have reported a pro-gramme of small scale tests to investigate three different adhesives, two ofwhich were two-part cold cure epoxies and the third a two-part acrylic. Thetests involved subjecting an adhesive/concrete joint to tensile force and acomposite/adhesive/concrete joint to shear, to verify the adequacy of thesurface preparation of the concrete and composite bond surfaces. In thesetests the Sikadur 31 PBA epoxy adhesive was superior to the two otherproducts and demonstrated strengths in both tension and shear whichexceeded those of the concrete. The acrylic adhesive failed within theadhesive under very small ultimate loads.

Chajes et al. (1996) used a single-lap specimen, in which a strip of carboncomposite was bonded to a concrete prism, to study the bond strength ofcomposite plate materials bonded to concrete. Four different adhesiveswere used to bond the composite strip; three two-part cold cure structuralepoxies and a two-part cold cure urethane. Three methods of surface prepa-ration were studied, varying in severity from untreated to mechanicallyabraded to expose the coarse aggregate. It was found that all epoxy-bondedjoints failed as a result of the concrete shearing directly beneath the bondsurface at similar loads. The final strength was therefore a function of theconcrete strength. The surface treatment which involved exposing thecoarse aggregate produced the highest average strengths. The urethaneadhesive, which was much less stiff and had a much higher ductility tofailure in tension than the epoxies, failed within the adhesive at lowerultimate loads. It is of interest to note that a silane surface primer was usedon two of Chajes’ adhesives (the primer used was Chemglaze 9926) and itimproved the bond performance of the joints compared with a joint nottreated thus; when used on concrete the primer tends to improve the bondby strengthening the surface of the concrete and making it water repellent.

Karbhari and Engineer (1996) describe the use of a modified peel test forinvestigation of the bond between composite and concrete, in which acomposite strip is pulled away from the concrete at a known angle and at acontrolled rate. The test is said to provide a good estimate of interfacialenergy and could be used in durability assessment.

SOR2 12/19/98, 2:39 PM16


2.3 External strengthening using steel plates

2.3.1 Introduction

A review of some significant experimental investigations conducted usingsteel plates is presented to demonstrate some of the structural implicationsof external plating. Research work into the performance of membersstrengthened with steel plates was pioneered simultaneously in SouthAfrica and France in the 1960s (L’Hermite and Bresson, 1967; Fleming andKing, 1967; Lerchenthal, 1967; Gilibert et al., 1976). Continued develop-ment of suitable adhesives and the increased use of the technique inpractice stimulated further research work. Eberline et al. (1988) presenta literature review on research and applications related to steel platebonding.

2.3.2 Structural investigations

The history of bonded external reinforcement in the UK goes back to 1975with the strengthening of the Quinton Bridges on the M5 motorway. Thisscheme followed a number of years of development work by the Transportand Road Research Laboratory (TRRL), (now TRL), in association withadhesive manufacturers and the Department of Transport. In terms oftesting programmes, research and development work continued at theTRRL and at several academic institutions in the UK, most notably at theUniversity of Sheffield. Theoretical investigations and the evaluation ofsuitable adhesives were allied to the extensive beam testing programmeswhich were undertaken.

Preliminary studies were conducted by Irwin (1975). Macdonald (1978)and Macdonald and Calder (1982) reported four point loading tests on steelplated RC beams of length 4900 mm. These beams were used to providedata for the proposed strengthening of the Quinton Bridges (Raithby, 1980and 1982), and incorporated two different epoxy adhesives, two plate thick-nesses of 10.0 mm and 6.5mm giving width-to-thickness (b/t) ratios of 14and 22, and a plate lap-joint at its centre.

In all cases it was found that failure of the beams occurred at one end byhorizontal shear in the concrete adjacent to the steel plate, commencing atthe plate end and resulting in sudden separation of the plate with theconcrete still attached, up to about mid-span. The external plate was foundto have a much more significant effect in terms of crack control and stiff-ness. The loads required to cause a crack width of 0.1 mm were increased by95%, whilst the deflections under this load were substantially reduced. Thepostcracking stiffness was found to be increased by between 35–105%depending upon the type of adhesive used and the plate dimensions.

SOR2 12/19/98, 2:39 PM17


The features of this work became the subject of a more detailed pro-gramme of research at the TRRL (Macdonald, 1982; Macdonald andCalder, 1982), in which a series of RC beams of length 3500mm were testedin four point bending. The beams were either plated as-cast or plated afterbeing loaded to produce a maximum crack width of 0.1mm. The effect ofwidening the plate whilst maintaining its cross-sectional area constant wasstudied. It was found that the plated as-cast and the precracked beams gavesimilar load/deflection curves, demonstrating the effectiveness of externalplating for strengthening purposes.

An extensive programme of research work carried out at the Universityof Sheffield since the late 1970s has highlighted a number of effects ofexternal, epoxy-bonded steel plates on the serviceability and ultimate loadbehaviour of RC beams. A brief summary of some of the research findingsis presented by Jones and Swamy (1995).

Steel plate strengthening of existing structures has also been investigatedin Switzerland at the Swiss Federal Laboratories for Material Testing andResearch (EMPA) (Ladner and Weder, 1981). Bending tests were carriedout on RC beams 3700mm in length, and the plate width-to-thickness (b/t)ratio was studied whilst maintaining the plate cross-sectional area constant.The external plate continued through and beyond the beam supports, withwhich they were not in contact, for a distance such that the bonded area(48000mm2) was the same for each plate width. The external plate was notbonded to the concrete beam except in the anchorage areas beyond thesupports. The results clearly showed that thin plating was more effectivethan thick narrow plating, as noted in studies conducted in the UK. Theeffective anchorage length la which allowed the plate to reach yield beforeshear failure adjacent to the bonded areas was found to be inversely pro-portional to the b/t ratio. Therefore, as b/t increased (wide, thin plates), theanchorage length decreased.

2.3.3 Plate separation and anchorage

The ultimate behaviour of steel plated RC beams appears to be closelyrelated to the geometry of the plated cross-section. For thin plates, failureusually occurs in flexure. However, if the plate aspect ratio falls below acertain value, separation of the plate from the beam can occur, initiatingfrom the plate end and resulting in the concrete cover being ripped off.These observations are consistent with the fact that simple elastic longitudi-nal shear stresses are inversely proportional to the plate width. Conse-quently, as the steel plate width decreases, the longitudinal shear stressesincrease. In addition, the bending stiffness of the plate increases, therebyincreasing the peeling stresses normal to the beam.

However, the levels of stress at the steel plate ends are thought to be well

SOR2 12/19/98, 2:39 PM18


in excess of those due to simple elastic considerations (Macdonald, 1982).Concentrations of shear and normal stress arise at the plate ends of beamssubjected to flexure as a result of stiffness incompatibility between the plateand concrete, which can only be accommodated by severe distortion of theadhesive layer. The sudden transition from the basic unplated members tothe plate reinforced member is usually situated in a region of high shear andlow bending moment. The changing bending moment and distortion in theadhesive layer causes a build-up of axial force at the end of the externalplate; this induces high bond stresses on the adhesive/plate and adhesive/concrete interfaces which may reach critical levels, thereby initiating fail-ure. The magnitude of these plate end stresses for externally strengthenedbeams depends upon the geometry of the plate reinforcement, the engi-neering properties of the adhesive and the shear strength of the originalconcrete beam (Swamy and Mukhopadhyaya, 1995). The existence of peakpeeling and shear stresses at the plate end, in addition to bending stresses,results in a biaxial tensile stress state which forces the crack initiated at theplate end to extend horizontally at the level of the internal steel.

When failure occurs in this way, the use of a more flexible adhesiveis advantageous, since the region over which the tensile strain builds upin the external steel plate is extended, thereby resulting in a lower peakstress. This has been verified experimentally by Jones et al. (1985), wherebeams strengthened using an adhesive with an elastic modulus of around1.0 � 103 Nmm�2 gave slightly improved strengths when failure occurredby plate separation than strengths given by an adhesive with a modulusof around 10 � 103 Nmm�2.

As the structural benefits of external plating with steel are enhanced bythe use of larger, thicker plates, an alternative to limiting the areas (orperhaps as a safeguard against separation), would be the provision of someform of plate anchorage. Jones et al. (1988) presented theoretical andexperimental studies into the problem of anchorage at the ends of steelplates. A series of RC beams 2500mm in length, strengthened with epoxy-bonded steel plates of 6.0 mm thickness were tested to investigate differentplate end anchorage schemes. Four 6.0mm diameter bolts at each end ofthe plate, which penetrated to a depth of 75mm, were used in one configu-ration, whilst different sizes of angle plates were also tried, one of whichcovered the extent of the shear span, and compared with those of a beamplated with a single unanchored steel plate of b/t ratio 21, which failedsuddenly by plate separation at a load which was below that of the unplatedcontrol beam. It was found that the anchorage detail had no apparent effecton the deflection performance of the beams. The use of bolts did notprevent debonding, but complete separation was avoided and increases instrength up to 8% over the unplated beam were achieved. The bondedanchor plates were more effective, producing yielding of the tensile plates

SOR2 12/19/98, 2:39 PM19


and allowing the full theoretical strength to be achieved, 36% above thatof the unplated beam. The anchorage detail was also found to affect theductility of the beams near the ultimate load. Unanchored, the beams failedsuddenly with little or no ductility. The beams with bolts or anchor plates allhad similar ductilities, at least as high as the unplated control.

Hussain et al. (1995) investigated the use of anchor bolts at the ends ofsteel plated beams, in an attempt to prevent brittle separation of the plate.In agreement with Jones et al. (1988) the bolts, which were 15mm indiameter and penetrated to half the depth of the beam, were found toimprove the ductility of the plated beams considerably, but to have only amarginal effect on the ultimate load. The percentage improvement in duc-tility due to the addition of bolts was found to decrease as the plate thick-ness increased. The end anchorage could not prevent premature failure ofthe beams, although in this case failure occurred as a result of diagonalshear cracks in the shear spans.

It will be realised that in providing anchorage to the steel plated beams,considerable extra site work is involved and this in turn will increase thecost of the plate bonding technique considerably. However, with steel platebonding this anchorage is completely necessary.

2.3.4 Disadvantages of external strengthening usingsteel plates

The in situ rehabilitation or upgrading of RC beams using bonded steelplates has been proven in the field to control flexural deformations andcrack widths, and to increase the load-carrying capacity of the memberunder service load for ultimate conditions. It is recognised to be an ef-fective, convenient and economic method of improving structural perfor-mance. However, although the technique has been shown to be successfulin practice, it also has disadvantages. Since the plates are not protected bythe concrete in the same way as the internal reinforcement, the possibilityof corrosion exists which could adversely affect the bond strength, leadingto failure of the strengthening system. Uncertainty remains regarding thedurability and the effects of corrosion. To minimise the possibility of corro-sion, all chloride-contaminated concrete should be removed prior to bond-ing and the plates must be subjected to careful surface preparation, storageand the application of resistant priming systems. After installation, theintegrity of the primer must be periodically checked, introducing a furthermaintenance task to the structure. The plates are generally prepared by gritblasting which, unless a minimum thickness of typically 6mm is imposed,can cause distortion.

Steel plates are difficult to shape in order to fit complex profiles. Inaddition, the weight of the plates makes them difficult to transport and

SOR2 12/19/98, 2:39 PM20


handle on site, particularly in areas of limited access, and can cause the deadweight of the structure to be increased significantly after installation. Elabo-rate and expensive falsework is required to maintain the steelwork inposition during bonding. The plates are required to be delivered to sitewithin flatness tolerances to prevent stresses being introduced normal to thebondline during cure. The weight of the plates and this flatness requirementgenerally restricts the maximum plate length to between 6–8 m. Since thespans requiring strengthening are often greater than this length, joints arerequired. Welding cannot be used in these cases since this would destroy theadhesive bond. Consequently, lapped butt joints have to be formed, addingfurther complications to the design of the system.

Studs are required to assist in supporting the steel plates during installa-tion and under service loading conditions. This is especially true towardsthe ends of the plates where anchorages are required due to the highbending stiffness of the plate. The position of these studs must therefore beestablished prior to bonding. This process can involve a considerableamount of site work. Finally, if the plates are loaded in compression, buck-ling may occur, causing the plates to become detached.

The process involved in strengthening with steel plates can therefore beconsidered as relatively time consuming and labour intensive.

2.4 External strengthening using composite

materials

2.4.1 Introduction

To overcome some of the shortcomings that are associated with steel platebonding, it was proposed in the mid-1980s that fibre reinforced polymer(FRP) plates could prove advantageous over steel plates in strengtheningapplications (Meier, 1987; Kaiser, 1989; Meier and Kaiser, 1991). Unlikesteel, FRPs are unaffected by electrochemical deterioration and can resistthe corrosive effects of acids, alkalis, salts and similar aggressive materialsunder a wide range of temperatures (Hollaway, 1993). Consequently,corrosion-resistant systems are not required, making preparation prior tobonding and maintenance after installation less arduous than for steel.

The reinforcing fibres can be introduced in a certain position, volumefraction and direction in the matrix to obtain maximum efficiency, allowingthe composites to be tailor made to suit the required shape and specifica-tion. The resulting materials are non-magnetic, non-conductive and havehigh specific strength and stiffness in the fibre direction at a fraction ofthe weight of steel. They are consequently easier to transport and handle,require less falsework, can be used in areas of limited access and do not addsignificant loads to the structure after installation. Continuous lengths of

SOR2 12/19/98, 2:39 PM21


FRP can be readily produced which, because of their low bending stiffness,can be delivered to site in rolls. The inclusion of joints during installation isthus avoided. With the exception of glass fibre composites, FRPs generallyexhibit excellent fatigue and creep properties and require less energy perkilogram to produce and transport than metals. As a result of easier instal-lation in comparison to steel, less site disruption should be experienced inthe process, allowing faster, more economical strengthening.

The benefits of utilising FRP materials over steel in plate bonding appli-cations are thus clear. The drawbacks are the intolerance to uneven bond-ing surfaces which may cause peeling of the plate, the possibility of brittlefailure modes (Swamy and Mukhopadhyaya, 1995) and the material cost,since fibre composites are between 4–20 times as expensive as steel in termsof unit volume. However, in a rehabilitation project, where material costsrarely exceed 20% of the overall project cost, the installation savings canoffset the higher material costs (Meier, 1992). Peshkam and Leeming (1994)have considered the commercial viability of FRP plate bonding for bridgestrengthening. In a straight comparison with steel plate bonding for a typi-cal application, despite the fact that material costs will be increased, labourand equipment costs will be reduced, construction times will be shorter anddurability will be improved. It is shown that 2kg of FRP could replace 47kgof steel on an equal strength basis. The costs of installing both materials areshown to be similar; however, when traffic management, traffic delay andmaintenance costs are included, the use of FRP provides a saving of 17.5%over steel. There are situations where steel plate bonding is not a viableoption because of the extent of chloride contamination of the concrete. Insuch cases, the use of FRP may avoid the need for demolition and replace-ment. Peshkam and Leeming (1994) presented a cost comparison of bridgereplacement against strengthening with FRP, in which possible savings of40% are demonstrated. These cost comparisons were made before truemanufacturing and installation costs were known and were at the bestestimates. Subsequently the tendering process for real installation projectshas shown carbon fibre reinforced polymer (CFRP) plate bonding to bevery competitive against steel plate bonding in first cost, before even futuremaintenance costs are added to the whole life cost equation.

The general versatility of composite materials makes them a viable alter-native to steel plates in strengthening applications, resulting in both shortterm and long term savings. Meier and Winistorfer (1995) consider that forapplications in which the possibility of corrosion is minimal and the lengthof the strengthening is less than 8m, steel will remain the most favourableoption. However, more recent work by other researchers (ROBUST) andtrends in costs, show that this position is changing and the indicationsare that FRP is more economical than steel whatever the length. This isthe case mainly in building construction, although plate thickness may be

SOR2 12/19/98, 2:39 PM22


important from an aesthetic viewpoint. In applications where corrosion,length of the required strengthening and handling on site are of greatersignificance, for example bridge rehabilitation, fibre composites become amore attractive alternative.

Concerns have been expressed regarding the behaviour of FRP strength-ened members when exposed to fire. A series of tests has been carried outat EMPA in Switzerland in which the performance of steel and CFRPplated beams was compared when exposed to extreme high temperatures(Deuring, 1994). It was found that a steel plate became detached after amatter of minutes of exposure, whereas the CFRP laminates progressivelylost cross-sectional area due to burning at the surface, causing a gradual lossof stiffness of the member, before final detachment after over an hour. Thissuperior behaviour is a consequence of the low thermal conductivity of thecomposite. In addition, detachment of a heavy steel plate from a structurefor any reason presents a far greater hazard than that of a lightweight FRCmaterial. Aspects of the effects of fire on resin compounds are consideredby Tabor (1978) and Hollaway (1993a).

Glass, aramid and carbon fibre composites may be considered forstrengthening applications. With particular regard to plate bonding, a com-parison of the important characteristics of FRP produced from these fibretypes is shown in Table 2.1, in which the fibre volume fraction is typicallyaround 65% and the fibres are unidirectionally aligned.

In common usage, glass is the most popular reinforcing fibre since it iseconomical to produce and widely available. However, concern exists re-garding the durability of composites composed of glass fibres, especially forstructural uses involving concrete, as discussed in Chapters 2 and 6. Carbonfibres exhibit better resistance to moisture, solvents, bases and weak acids,and can withstand direct contact with concrete (Santoh et al., 1983). Com-posite materials produced from them are light in weight, with strengths

Table 2.1 Comparison of characteristics of FRC sheet produced from differentfibres (Meier, 1995)

Characteristics Carbon Aramid E-glass

Tensile strength Very good Very good Very goodCompressive strength Very good Inadequate GoodStiffness Very good Good AdequateLong term behaviour Very good Good AdequateFatigue behaviour Excellent Good AdequateBulk density Good Excellent AdequateAlkaline resistance Very good Good InadequateCost Adequate Adequate Very good

SOR2 12/19/98, 2:39 PM23


higher than steel and stiffnesses higher than either glass or aramid compos-ites. For example, laminates fabricated from glass fibre must be three timesthicker than CFRP laminates to achieve the same tensile stiffness for thesame fibre volume fraction. CFRP has excellent fatigue properties and avery low (or even negative) linear thermal coefficient of thermal expansionin the fibre direction. Quality assurance can be performed by non-destructive testing, for example infrared inspection in the field, if CFRPlaminates are used; this is not possible with steel plates. This techniqueallows fast and accurate judgement on the quality of the strengthening work.

Despite the higher cost, carbon composites appear to provide the bestcharacteristics for structural strengthening.

2.4.2 Review of experimental investigations

The following section reviews, on a geographical basis, experimental workreported to investigate the flexural strengthening of RC members usingnon-prestressed FRP plates. These studies have utilised fibrous materials invarious forms, including pultruded plates, precured prepreg plates, prepregsheets or tapes cold laminated in place, and dry fibre sheets impregnated atthe time of bonding.

2.4.2.1 Some investigations in Europe

Recent work on the use of FRP materials as a replacement for steel in platebonding applications was pioneered at the EMPA in Switzerland. Fourpoint loading tests were initially performed on RC beams 2000mm (Meier,1987; Kaiser, 1989) or 7000mm (Ladner et al., 1990) in length. Strengthen-ing was achieved through the use of pultruded carbon fibre/epoxy laminatesup to 1.0mm thick bonded with the same epoxy adhesives used in earliersteel plating work (Ladner and Weder, 1981). For the 2000mm lengthbeams, the ultimate load was almost doubled over the unplated controlbeam, although these beams were designed with a low proportion of inter-nal steel, and hence the strength of the unplated beam was low. In the caseof the 7000 mm length beam, strengthened with a 1.0mm CFRP laminate,the increase in the ultimate load was about 22% (Ladner and Holtgreve,1989). However, for both beam lengths the ultimate deflection was con-siderably reduced, although it was claimed that there was still sufficientrotation to predict impending failure.

The following modes were observed either individually or in combinationin the tests carried out at the EMPA:

• sudden, explosive, tensile failure of the CFRP laminates• compressive failure in the concrete

SOR2 12/19/98, 2:40 PM24


• slow, continuous peeling of the laminate during loading resulting froman uneven concrete bond surface

• sudden peeling of the laminate during loading due to relative verticaldisplacement across a shear crack in the concrete

• horizontal shearing of the concrete in the tensile zone• interlaminar shear within the CFRP sheet.

The CFRP plate was found to reduce the total width of cracks andproduce a more even crack distribution over the length of the beam (Meierand Kaiser, 1991). Meier et al. (1992) recommended that in strengtheningapplications, the external CFRP should fail in tension after yielding theinternal steel but before failure of the concrete in the compressive zone,since this would ensure a more ductile failure mode.

Deblois et al. (1992) investigated the application of unidirectional andbidirectional glass fibre reinforced polymer (GFRP) sheets for flexuralstrengthening. A series of RC beams 1000mm long were tested afterstrengthening. The use of bidirectional sheets increased the ultimate loadby up to 34%, whereas unidirectional GFRP resulted in an increase of only18%. The authors of this current chapter feel that this is an unexpectedconclusion and emphasise that the FRP material used was GFRP. Theadditional bonding of bidirectional GFRP to the sides of the beam in-creased the load carried with unidirectional sheets to 58%. To further theprogramme of study, Deblois et al. (1992) epoxy-bonded a bidirectionalGFRP sheet to the soffit of a 4100mm long RC beam. Bolts were also usedas additional anchorage at the plate ends. The maximum load increased by66% over the unplated control beam. It was noted that for all tests theapplication of GFRP reduced the ductility of the beam.

Research into external FRP plating has been conducted at OxfordBrookes University (Hutchinson and Rahimi, 1993). The effect of plate-endgeometries on the stress concentrations at the plate ends was of primaryinterest in this investigation, for which RC beams 2300mm in length wereused. Two beams were preloaded to 80% of their ultimate strength, beforeplating, to cause cracking of the concrete and yielding of the steel reinforce-ment. Unidirectional carbon fibre/epoxy prepreg tape of total thickness0.78 mm was used for the plating, with the various plate end geometries,tapering in either plan or section, cut whilst the composite was in theuncured state. Several different two-part cold cure epoxy adhesives wereevaluated using a modified Boeing wedge cleavage test, of which no detailis given, developed to measure adhesion to concrete surfaces. The adhesiveselected as most suitable was Sikadur 31 PBA, an epoxy which has beenused in both steel and FRP plate bonding applications in Switzerland(Meier and Kaiser, 1991) and in steel plating applications in the UK (Shaw,1993).

SOR2 12/19/98, 2:40 PM25


It was found that the flexural performance of all strengthened beams wassignificantly better than the unplated specimens, in terms of both strengthand stiffness. The ultimate load-carrying capacity was increased by as muchas 230%; however, it should be pointed out that the actual increase isdependent upon the degree of internal reinforcement in the beam beforeplating. The increased stiffness resulted in an increased load to first crack-ing, but a substantial decrease in ductility to failure. After first cracking,cracks grew progressively in number, covering most of the test span. Mostof these were of hairline width even close to the ultimate load level. In allcases, failure was sudden and catastrophic, characterised by a shear crackrunning from the tensile zone towards the loading point and delaminationof the concrete cover along the tensile reinforcement. This type of failurehas been identified in steel plating work as described above. Tapering of theplate end in either plan or section appeared to have no effect on the flexuralperformance or failure mode for the cases considered.

As with steel plates, the beams which had been precracked before bond-ing had an equivalent performance to the other test beams, indicating theeffectiveness of the plate bonding technique for repair. The load/deflectionbehaviour was similar for all different plate configurations, except for thosewith laminates bonded to the full length of the beam, clamped by thereaction at the supports, which resulted in an increase in strength over theother plated beams. It was concluded that for these particular beams andplates the ultimate loading capacity of the system appeared to have beenreached, being governed by the shear capacity of the concrete beams.

The tests at Oxford Brookes University continued (Hutchinson andRahimi, 1996), under the ROBUST programme of research, by utilisingboth glass and carbon fibre/epoxy laminates of different thicknesses built upfrom prepreg tapes. Three internal steel reinforcement ratios were exam-ined. All beams with external reinforcement performed significantly betterthan their unplated counterparts in terms of stiffness and strength. The useof GFRP was found to provide significant ductility and reasonable strength,whilst enhancements were greater with CFRP but at the expense of a loss ofductility. Greater enhancements were achieved with lower steel ratios.

A limited programme of experimental testing has been carried out at theUniversity of Bologna (Arduini et al., 1994; Arduini et al., 1995) in whichsmall scale steel fibre reinforced concrete specimens of length 500mm or600mm have been tested in three point bending after being strengthenedwith unidirectional aramid fibre/epoxy or glass fibre/epoxy composites ofthickness between 2.0–5.0mm. These small scale tests were used to demon-strate that the load-carrying capacity of the basic unplated beam could beincreased through external plating with FRC but that different failuremodes, often brittle, were involved. It was noted that peeling and shearcracks at the plate ends were responsible for causing premature, brittle

SOR2 12/19/98, 2:40 PM26


failure. The use of thicker FRC plates was found to increase the occurrenceof peeling failure. Ductility was increased and peeling failure delayedthrough the use of plates bonded to the sides of the beams in the plate endregions; the effects were enhanced by coupling the side and soffit platestogether, in which case failure was observed to occur by diagonal shearingat the highest attained loads.

The University of Surrey (Quantrill et al., 1995) under the ROBUSTprogramme of research, undertook a parametric study on flexurallystrengthened RC beams using GFRP bonded plates. The study involvedvarying the concrete strength, the pultruded composite plate area and itsaspect ratio (b/t), and as discussed above in steel plating applications, thicknarrow plates with aspect ratios of less than 50 have been associated withbrittle peeling failure modes. Consequently, ratios of 38 and 67 were testedin the study. The effect of the b/t ratio was isolated in these tests bymaintaining a constant plate cross-sectional area. The tests showed thatplating can considerably enhance both the strength and stiffness of RCmembers, although at the expense of ductility at failure. It appeared thatthe higher strength concrete produced the greatest increase in strength overthe unplated section and the aspect ratio of the plate has little effect on theoverall behaviour.

The above programme continued with further investigations at theUniversity of Surrey (Quantrill et al., 1996a) into the experimental andanalytical strengthening of reinforced concrete beams with fibre reinforcedpolymer plates, and analysed the effects of different plate parameters onthe overall behaviour of the system. It was shown that testing relativelysmall scale 1 m long specimens can reveal useful information on strength-ened beam behaviour. By reducing the plate area the expected reductionin strengthening and stiffening caused the ductility and the plate strains fora given load to increase; the aspect ratio for the values tested had littleeffect on the overall response. Plating with CFRP components increasedthe serviceability, yield and ultimate loads and increased the strengthenedmember stiffness after both cracking and yielding; ductility was reduced.The iterative analytical model accurately predicted the tensile plate strainand compressive concrete strain responses of the beam for a partiallycracked section.

Quantrill et al. (1996b) continued with tests on small scale specimens andshowed that when the CFRP plated beams were uncracked at their extremi-ties the theoretical shear stress reached 11.15N mm�2 and the peel stress6.37 N mm�2. The anchored CFRP plated beams were able to sustain higherlevels of shear and peel stress before failure occurred around 14.1Nmm�2 inshear and 8.10 N mm�2 in peel stress.

In France, a programme of small scale tests has also been carried out tostudy the effects of different adhesive and FRC combinations when used for

SOR2 12/19/98, 2:40 PM27


external strengthening (Varastehpour and Hamelin, 1995). A series of plainconcrete specimens 280mm in length were tested to failure in four pointbending after being strengthened with glass or carbon fibre/epoxy sheetsbonded to both the tension and side faces of the specimen with one of fourdifferent epoxy or acrylic adhesives. The composite plate on the tensionface was anchored by the reactions at the supports in all cases. Failureoccurred either by FRP rupture, interface failure or by debonding of theplate from the concrete. In all cases the flexural and shear capacity of thebeams was increased by plating, although this was found to be dependenton the choice of adhesive; in general, the epoxies performed better thanthe acrylics, the tests demonstrating that a rubber-toughened epoxy wassuperior.

Under the ROBUST research programme, Garden et al. (1996) showedthat the ultimate capacity of the CFRP beams falls with reducing the width–thickness b/t and beam shear span/depth ratios. Failure under low shearspan/beam depth ratios is associated with high plate strains (the value beingin the region of 70% of the plate ultimate strain) and relatively high longi-tudinal shear stresses at the adhesive/concrete interface, and although theconcrete failed in the cover concrete area, debonding from the concrete wasnot observed. Plate end anchorage delays failure by resisting plate separa-tion but does not increase stiffness until the internal reinforcement hasyielded.

He et al. (1997a), at the University of Sheffield, used steel and CFRPplates with the same axial stiffness-to-strength precracked reinforced con-crete beams in which a new, but unspecified, plate anchorage system wasadopted. The basic improvement in structural performance due to platingwas verified and it was found that the CFRP plates produced a greaterimprovement in ultimate load than the steel plates. The authors (He et al.,1997b) noted that the high stress and strain potential of the CFRP will notbe utilised unless the plate is prestressed.

Bencardino et al. (1997) tested CFRP plated beams at the University ofCalabria, Italy, recording reductions in member ductility due to platingwithout end anchorage; the ductility was restored when anchorage wasfitted in the form of externally bonded U-shaped steel stirrups. The methodof CFRP plating was used successfully to strengthen an experimental portalstructure.

2.4.2.2 Some investigations in North America

During the late 1980s, a pilot study was carried out at the University ofArizona to establish the feasibility of poststrengthening concrete bridgebeams with GFRP plates (Saadatmanesh and Ehsani, 1989 and 1990a).Selection of a suitable epoxy adhesive for plate bonding purposes was the

SOR2 12/19/98, 2:40 PM28


main subject of investigation. Five RC beams 1675mm in length were testedin four point bending to determine their static strength. None of the beamscontained shear reinforcement, which resulted in premature failure in thefirst tests. To prevent shear cracks causing separation of the plate from thebeam, external shear reinforcement was thus provided in the end regionsby means of several large G-clamps. One beam was unplated, while theremainder were strengthened with a 6.0 mm thick GFRP plate bonded withone of four different types of two-part cold cure epoxy with a range of shearstrengths from 13 MPa to 16MPa using aluminium substrates.

It was found that flexible epoxies did not allow any measurable shear tobe transferred between the plate and the beam, and no increase in theultimate strength was achieved in comparison to the unplated control beam.For the most rigid epoxy, after the concrete had cracked in tension, theplate was observed to separate from the beam in a very brittle manner,again resulting in no increase in ultimate capacity. The beam strengthenedusing a relatively viscous rubber-toughened epoxy was found to performbest in the tests carried out. This beam was significantly stronger and stifferthan the unplated control beam. Substantial force developed in the plateindicated good shear transfer and composite action between the plate andthe concrete beam. The cracks were found to be considerably smallerthroughout the range of loading and distributed more evenly along thelength of the beam. Failure occurred when a layer of concrete delaminatedabout 10 mm above the bondline, indicating satisfactory performance of theepoxy.

Following on from this initial study, a further experimental researchproject was undertaken at the University of Arizona (Saadatmanesh andEhsani, 1990b and 1991). In this project, five rectangular beams and oneT-beam were tested to failure in four point bending over a clear spanof 4570 mm. All beams were strengthened with a GFRP plate 6.0mmthick bonded to the concrete with the epoxy adhesive identified as mostsuitable for the application from the previous study. Three different rein-forcement ratios were used for the tension steel in the beams. The majorityof the beams were overdesigned for shear to prevent premature shearfailure.

The tests indicated that significant increases in the external load, at whichthe steel yielded, and an increase in flexural strength, could be achieved bybonding GFRP plates to the tension face of RC beams. The gain in theultimate flexural strength was found to be significant in beams with lowersteel reinforcement ratios, as noted later by Hutchinson and Rahimi (1996).In addition, plating reduced the crack size in the beams at all load levels.For several of the beams tested, failure occurred as a result of suddenlongitudinal shear failure of the concrete between the plate and internalsteel reinforcement. The flexural stiffness was increased, although the

SOR2 12/19/98, 2:40 PM29


ductility of the beams and curvature at failure were reduced by the additionof the GFRP plates.

An experimental programme was undertaken by Chajes et al. (1994) atthe University of Delaware, in which RC beams 1120mm in length wereloaded to failure in four point bending after the majority had been exter-nally strengthened with composite fabric of either bidirectional wovingaramid, E-glass or carbon fibre reinforcement. In each case, the fabrics hada tensile capacity close to the yield strength of the steel. Fabrics were usedas an alternative to plates to exploit their ability to conform to irregularsurface geometries, thus reducing the possibility of the continuous peelingfailures observed in testing at the EPMA. No shear reinforcement wasprovided in the beams. A variety of layers of each fabric were epoxy-bonded to the concrete. A set of three beams were also prepared and testedwith twice the amount of internal steel reinforcement.

It was found that the mode of failure of the strengthened beams varieddepending upon the fabric used; those externally strengthened with E-glassand carbon fibres failed by tensile rupture of the fabric. The first aramidstrengthened beam exhibited a fabric debonding mode of failure and conse-quently, for the remaining specimens, end tabs were included, bonded tothe sides of the beam and enclosing the soffit reinforcement in the endregions. The extent of the end tabs in the shear spans is not clear from thepublication, but their use prevented debonding, allowing failure of theconcrete in compression to occur.

For each of the fabric types used, increases in flexural strength similarto those found in the beams with additional steel reinforcement wereachieved, in the range 34–57%. The fabric reinforcement beams also exhib-ited increases in flexural stiffness within the range 45–53%. Both of thefailure modes observed were said to yield a reasonable amount of ductility,although this was around half that obtained from the unstrengthenedbeams.

The research carried out at the University of Delaware forms part of awider study concerned with the possibility of rehabilitating deterioratedprestressed concrete box beam bridges using transversely bonded advancedcomposite materials (Chajes et al., 1993; Finch et al., 1994; Chajes et al.,1995b; Chajes et al., 1996).

The US Navy has been studying the possibility of using external FRPplating for upgrading waterfront structures affected by reinforcement cor-rosion (Malvar et al., 1995). Enhancements of both bending and shearstrength are being considered through the use of unidirectional CFRP towsheets. RC beams 1680mm long have been tested in an experimental inves-tigation, none of which contained shear reinforcement. Beams strength-ened longitudinally demonstrated that the flexural strength could besignificantly enhanced, but failure occurred, not surprisingly, in shear.

SOR2 12/19/98, 2:40 PM30


When additional CFRP was wrapped onto the sides and soffit of the beamover its full span, to provide shear reinforcement and additional anchoragefor the longitudinal CFRP sheets, sufficient shear strength was provided torevert to a bending failure in which the steel yielded, the concrete crushedand then the CFRP material ruptured. However, this occurred at a ductilitywhich was somewhat less than that of the unplated control beam.

In addition to upgrading reinforced concrete beams, research into thefeasibility of externally reinforcing continuous RC slab bridges in responseto observed longitudinal cracking was initiated in South Dakota (Iyer et al.,1989). To close the observed cracks, the possibility of bonding the externalreinforcement whilst the beam was relieved of dead load was examined.The use of both steel (Iyer et al., 1989) and CFRP plates (Iyer, 1988) hasbeen reported. Initial results on small scale beams showed that the strains inthe concrete and internal steel were considerably reduced by the introduc-tion of external reinforcement, while the stiffness was increased and crack-ing was controlled.

2.4.3 Prestressing composite plates for strengtheningconcrete beams

The utilisation of prestressed composite plates, at the time of bonding, forstrengthening concrete members has been studied only relatively recentlyin comparison with investigations of non-prestressed plates, although thebenefits of external prestressing with plate materials have been recognisedfor many years. For example, Peterson (1965) considered the externalprestressing of timber beams using prestressed steel sheets and found sig-nificant improvements in bending stiffness and ultimate capacity. Externalprestressing with composite plates also provides these benefits as well ascost savings. Triantafillou and Deskovic (1991) noted that this method ofprestressing is a more economical alternative to conventional prestressingmethods used in new construction.

Initial research on the strengthening of reinforced concrete beams byexternal plate prestressing at EMPA in Switzerland has been widely re-ported (Meier and Kaiser, 1991; Meier et al., 1993; Deuring, 1994). Thiswork included the cyclic loading of a beam whose plate was prestressed to50% of its strength. Although this prestress ensures the mean stress level inthe cyclic loading was high, there was no evidence of damage to the plateafter 30 � 107 cycles and the cracking of the concrete was well controlled.The non-prestressed beam loading tests reported by Deuring (1993) re-vealed failures by the initiation of plate separation from the base of a shearcrack. It was found that the compression transfer into the concrete by theplate prestress could delay or even prevent this type of failure, therebyallowing the plate to reach its ultimate tensile strain so that the beam failed

SOR2 12/19/98, 2:40 PM31


in flexure rather than by premature plate separation (Deuring, 1993). Theability of the plate to alter the failure mode from premature plate separa-tion to flexure is influenced by the prestressing force and the cross-sectionalarea of the plate. One of the conclusions of the work was that the greatestflexural resistance of a strengthened section is reached when the platefractures in tension, either after or at the same time as yield of the internalsteel rebars.

Saadatmanesh and Ehsani (1991) conducted an experimental study of thestrengthening of reinforced concrete beams using non-prestressed and pre-stressed GFRP plates. One of the two prestressed beams contained a rela-tively small amount of internal tensile steel reinforcement, while the othercontained larger bars and was precracked prior to bonding of the plate. Theplate prestress in the precracked case closed some of the cracks, indicatingthe benefit of prestressing from a serviceability point of view. The beamwith little original reinforcement before plating experienced a large im-provement in ultimate capacity due to the additional moment couple pro-vided by the plate prestress. In both cases, the prestress was generated bycambering the beam before bonding the plate so that a tensile load wastransferred to the plate when the camber was released. Improved con-crete crack control was observed with prestressed plates, a clear advantagefrom a serviceability point of view. A previous experimental study bySaadatmanesh and Ehsani (1990) also included the prestressing of a beamby cambering; this particular specimen had a low internal reinforcementarea ratio of 0.32% (based on effective depth) so that a very high increasein ultimate load (323%) was observed as a result. The GFRP plate was notanchored at its ends, and failure occurred by premature plate separationassociated with the removal of a layer of concrete from the tension face ofthe beam.

Triantafillou et al. (1992) tested reinforced concrete beams in three pointbending with various quantities of internal reinforcement and magnitudesof CFRP plate prestress. Improved control of concrete cracking wasbrought about not only by a greater internal reinforcement provision, butalso by higher plate prestress, indicating the serviceability advantage gainedby prestressing the composite. It was noted that prestressed compositeplates can potentially act as the sole tensile reinforcement in new concreteconstruction and prefabrication is also possible due to the simplicitywith which composites may be handled and applied. The confinementimposed by the initial compressive stress at the base of the beam wasthought to be capable of improving the shear resistance of the member.Also, an advantage from a cost point of view is that the same strengtheningto failure may be achieved with a prestressed plate of relatively small cross-section, like that achieved with a larger non-prestressed plate (Triantafillouet al., 1992).

SOR2 12/19/98, 2:40 PM32


Char et al. (1994) conducted an analytical parameter study to determinethe effects of varying the cross-sectional area and material type of thecomposite plate and the prestress in the plate. The parameter study re-vealed that prestressing a GFRP plate would not necessarily increase theultimate moment capacity over that of a beam with a non-prestressed plate,for the particular beam configuration and prestress level considered. Thiswas because both the non-prestressed and prestressed beams failed by platefracture. Garden and Hollaway (1997) showed that prestressing with CFRPplates increases the ultimate capacity of a beam but the magnitude of theincrease depends on the failure modes of the beams with and withoutprestress; the failure mode of the prestressed beam depends on the pre-stress magnitude.

Wight et al. (1995) reported data on the strengthening and stiffeningachieved with prestressed CFRP plates. The control of concrete crackwidths and numbers of cracks was improved by prestressing the plates. Thebeam with an initial non-stressed plate failed by concrete fracture in thecover thickness within one of the shear spans of the four point loaded beam,whereas the prestressed plated beams failed by plate fracture in the con-stant moment region. The compression generated in the concrete nearthe beam soffit, due to the plate prestress, was sufficient to reduce themagnitudes of vertical displacements across shear cracks and to transferfailure into the plate. The avoidance of concrete failure in the shear spanswas associated with a much improved ultimate load.

The testing at the University of Surrey, under the ROBUST programme,continued (Quantrill and Hollaway 1998) by pretensioning the ROBUSTpultruded composite plates, prior to bonding to the concrete. The pre-stressing technique employed was developed and refined on small scale1.0 m long specimens before being applied to larger 2.3m long beams.Pretensioning the plate prior to bonding to the concrete beam considerablyincreased the external applied load at which cracking of the concrete oc-curred, reduced overall member stiffness and also the load at which visiblecracking occurred. The observation of crack control is of significant impor-tance to serviceability based design criteria. It was generally concludedthat this technique has the potential to provide a more efficient solution tostrengthening problems.

Furthermore, Garden and Hollaway (1997) tested 1.0 m and 4.5m lengthsof reinforced concrete beams in four point bending after strengtheningthem with externally bonded prestressed CFRP plates. The plates werebonded without prestress and with prestress levels ranging from 25–50% ofthe plate strength. The ultimate capacities of the plated non-prestressedbeams were significantly higher than those of the unplated members andplate prestress brought about further strengthening. The non-prestressedbeams failed by concrete fracture in the cover to the internal rebars, whilst

SOR2 12/19/98, 2:40 PM33


most of the prestressed beams failed by plate fracture. The plate prestressprevented cracking of the adhesive layer, a phenomenon associated withshear cracking in the concrete. The plates of the prestressed beams had aninitial tensile strain before any external load was applied to the beamsystem and consequently at this stage, the beams had a relatively highstiffness. It was found that prestressed plates were utilised more efficientlythan non-prestressed plates since a given plate strain was associated with alower plated beam deformation in a prestressed member. Prestressing thecomposite plates lowers the position of the neutral axis so more of theconcrete section is loaded in compression, making more efficient use ofthe concrete.

All the above experiments were carried out in the laboratory on rela-tively small scale beams and the method of prestressing could not have beenused on site on a real structure where the plate would have to be stressedbefore bonding within the confines of the abutments or supports. Within theROBUST project, two 18m long beams recovered from a real bridge struc-ture, which had to be demolished and reconstructed, were strengthenedwith plates that were prestressed under conditions that were little differentfrom those that would occur on a real bridge structure (Lane et al., 1997).These tests are described in more detail in Chapter 5.

2.5 Strengthening of reinforced concrete

members in shear

Some research work has been conducted on the use of fibre reinforcedcomposite plates for strengthening structures in shear.

Al-Sulaimani et al. (1994) experimentally studied the use of GFRP platesfor the shear strengthening of initially shear-cracked concrete with a shearcapacity 1.5 times lower than their flexural capacity. A low shear span/beamdepth ratio of 2.7 was used, which would have ensured that shear wasdominant in the beam behaviour. The shear repair comprised three differ-ent systems, with and without the soffit plate in each case. The first repairinvolved the external bonding of 20mm wide strips over the side and soffitof the beam at regular intervals throughout each shear span. The secondrepair utilised the bonding of side plates, throughout each shear span,covering 80% of the beam depth and located centrally in the depth. Thethird method involved the bonding of a U-shaped jacket covering the sidesof the beam and the soffit plate throughout each shear span. The beamsrepaired with side strips and side plates failed by diagonal tension, withdominant cracks at failure following the cracks initially present in thebeams from the preloading stage. Concrete compression failure occurred inthe beams with the jackets.

The programme of experimental work by Chajes et al. (1994), on small

SOR2 12/19/98, 2:40 PM34


scale specimens, concentrated on GFRP composites as the external rein-forcing medium (Chajes et al., 1995a). Increases in flexural and shear capac-ity of beams 1120 mm in length were examined when tested to failure infour point bending. These small scale beams, which again had no shearreinforcement, were externally strengthened with unidirectional CFRP towsheets to the basic control beam configuration. To evaluate the effect ofcomposite shear reinforcement, a CFRP sheet was wrapped around thesection; again, the extent of this reinforcement along the span is unclear. Itwas found that the control beam was increased by 158% by adding a singleCFRP sheet to the tensile face of the beam. Increases in the load crackingof the concrete and yielding of the internal steel were also noted. In addi-tion to the increase in capacity, a 115% increase in stiffness, a changein failure mode from flexural to shear, and a decrease in ductility wereobserved. By wrapping the beam with a CFRP sheet, shear failure wasprevented and tensile failure of the composite occurred. Finally, by addinga second CFRP sheet to the tensile face, a 292% increase in capacity and a178% increase in stiffness were achieved. It should be stressed, however,that these large percentages are a function of the initial structural capacitiesof the beam.

Chajes et al. (1995b) tested beams reinforced externally with CFRPplates bonded to their soffit and sides to study flexural and shear behav-iours. The fibre orientation in the shear plates was in the vertical directionof the beams only. This orientation was believed to be the reason for thesimilarity in the load–deflection responses of flexurally strengthened beamswith and without external material; the vertical fibres had little effect on theflexural behaviour of the beams. The composite material used by Chajes etal. (1995b) was a unidirectional CFRP tow sheet having a dry thickness of0.11 mm and a tensile modulus of elasticity of 227.37GPa. The continuousstrips were able to control shear crack opening due to their greater axialstiffness, resulting in reduced shear deflection. This result showed that,unlike the flexural soffit reinforcement, a thin sheet covering as much of theconcrete as possible will not necessarily produce the greatest improvementin crack control where shear is concerned, but the sheets were able to avoidconcrete shear failure, the failure mode observed without the sheets. Thetests showed a logical progression of failure modes as more and moreexternal reinforcement was added. There was an increase in capacity of115% in stiffness, a change in failure mode from flexure to shear and adecrease in ductility. When a further single CFRP sheet was applied to thebeam, shear failure was prevented and a flexural failure initiated as a tensilefailure of the composite occurred. Finally by adding a second single layer ofCFRP sheet to the tensile face a 292% increase in capacity and a 178%increase in stiffness were achieved.

Taljsten (1997) studied the shear force capacity of beams when these had

SOR2 12/19/98, 2:40 PM35


been strengthened by CFRP composites applied to the beams by fourdifferent techniques. These were:

• hand-lay-up, by two different systems• prepreg in combination with vacuum and heat• vacuum injection.

The results of the four point loaded tests showed, in all cases, a goodstrengthening effect in shear when the CFRP composites were bonded tothe vertical faces of concrete beams. The strengthening effect of almost300% was achieved and it was possible to reach a value of 100% with aninitially completely fractured beam. Generally it was easier to apply thehand-lay-up system and Taljstenn suggested that although the prepreg andvacuum injection methods gave higher material properties than those of thehand-lay method, the site application technique seemed to be more con-trollable for the hand-lay process.

Hutchinson et al. (1997) has described tests that were undertaken at theUniversity of Manitoba to investigate the shear strengthening of scaledmodels of the Maryland bridge which required shear capacity upgrading inorder to carry increased truck loads. The bridge had an arrangement ofstirrups which caused spalling off of the concrete cover followed by straight-ening of the stirrups and sudden failure. CFRP sheets were effective inreducing the tensile force in the stirrups under the same applied shear load.The CFRP plates were clamped to the web of the Tee beams in order tocontrol the outward force in the stirrups within the shear span. This allowedthe stirrups to yield and to contribute to a 27% increase in the ultimateshear capacity. Hutchinson showed that diagonal CFRP sheets are moreefficient than the horizontal and vertical CFRP sheet combination in reduc-ing the tensile force in the stirrups at the same level of applied shear load.

2.6 Applications of FRP strengthening

Of the applications of FRP strengthening reported in the literature, themajority occur in Switzerland where the concept was first proposed anddeveloped. In these cases, which are considered in more detail by Meier(1995), pultruded carbon fibre/epoxy laminates have been used exclusively.The first reported application was the repair in 1991 of the Ibach Bridge inthe canton of Lucerne, for which several prestressing tendons had beensevered during the installation of traffic signals. The bridge was repairedwith three CFRP sheets of dimensions 150 mm wide by 5000 mm long andof thickness 1.75 mm or 2.0 mm. The total weight of the CFRP used wasonly 6.2 kg, compared with the 175 kg of steel which would have beenrequired for the repair. In addition, all work was carried out from a mobileplatform, eliminating the need for expensive scaffolding. A loading test

SOR2 12/19/98, 2:40 PM36


with an 840 kN vehicle demonstrated that the rehabilitation work had beensatisfactory.

The wooden bridge at Sins in Switzerland was stiffened in 1992 to meetincreased traffic loading (Meier et al., 1993). Two of the most highly loadedcross beams were strengthened using 1.0mm thick CFRP laminates. Theappearance of the historic structure was unaltered by the strengtheningtechnique. Other CFRP strengthening applications in Switzerland includeslab reinforcement around a newly installed lift shaft in the City Hall ofGossau St. Gall, the upgrading of a supermarket roof using laminates 15.5min length to allow the removal of a supporting wall, ground floor strength-ening of the Rail Terminal in Zurich, and the strengthening of a multistoreycar park in Flims. A chimney wall at the nuclear power plant in Leibstadthas also been poststrengthened for wind and seismic loading after theinstallation of ducts.

Rostasy et al. (1992) report the use of GFRP plates at the working jointsof the continuous multispan box girder Kattenbusch Bridge in Germany toreduce fatigue stresses in the prestressing tendons and transverse crackingdue to thermal restraint. A representative specimen of the joint was testedin the laboratory to verify the technique prior to field application. Ten jointsrequired rehabilitation; eight of these were strengthened with steel plates10 mm thick, whilst the remaining two utilised GFRP plates 30mm thick toprovide the same area stiffness as the steel plates. The installation of suchplates, of which twenty were used at each joint, took place in 1987 and wasfound to reduce the stress amplitude at the joints by 36% and the crackwidths by around 50%.

Greenfield (1995) describes applications of composite strengthening inthe United States, in which the integrity of a sewage treatment basin wasrestored with carbon fibre/epoxy laminates 1.65mm thick. The laminateswere also used to relieve overstress in areas of the basin due to lack ofreinforcing steel. The seismic retrofit of bridge columns in California usingGFRP jackets has been reviewed by Priestley et al. (1992).

An existing roof structure at Kings College Hospital, South London hasbeen strengthened using epoxy-bonded, 1.0 mm thick, 11m long pultrudedCFRP laminates (NCE, 1996). An extra floor was added to the buildingsuch that the existing roof was strengthened to meet new floor require-ments. The installation took place quickly and conveniently, 2 kg of CFRPbeing used instead of 60kg of steel.

Nanni (1995) reported the findings of a visit to Japan to determine thescale of FRC use as external reinforcement. He concluded that a greaternumber of field applications in Japan in recent years have used thinner FRCsheets than the plates used in Europe, Saudi Arabia and North America.The use of FRC sheets for the structural strengthening of concrete in Japanhas addressed problems in bridges, tunnels, car parks and other structures

SOR2 12/19/98, 2:40 PM37


(Greenfield, 1995). The following five examples of FRC strengthening werecited by Nanni (1995), carbon fibre composites having been used in allcases:

• Strengthening of a cantilever slab of the Hata Bridge along the KyushuHighway in order to accommodate large parapet walls which causedelevated bending moments due to the higher wind force;

• Increase of the load rating of the Tokando Highway bridge atHiyoshikura, a reinforced concrete deck supported on steel girders,causing a 30–40% reduction of stress in the internal rebars;

• Arrest of the internal steel reinforcement corrosion of the concretebeams in the waterfront pier at the Wakayama oil refinery;

• Strengthening and stiffening of the concrete lining of the Yoshino Routetunnels on Kyushu Island, necessary due to cracking which arose fromunexpected fluctuations in the underground water pressure. No loss oftunnel cross-sectional area occurred and the road remained open duringthe bonding work;

• Longitudinal strengthening of the sides and soffit of a culvert at theFujimi Bridge in Tokyo.

Chapter 11 of this book discusses some case histories of plate bondingundertaken in the UK and in Europe and goes into the technologies used inthose cases in greater depth than has been possible in this chapter.

2.7 Summary and conclusions of literature review

The review given in this chapter is based on steel and composite platebonding and has been covered extensively but not exhaustively. It hasdemonstrated the improvement in structural strength and stiffness broughtabout by externally bonded material. The worldwide level of interest in thetechnique reflects its potential benefits and also the current importanceplaced on economical rehabilitation and upgrading methods. Although thelevel of experience in the bonding technique of composite plates is limited,the investigations reported in this chapter have gone some way to illustrateits potential and to establish a basic technical understanding of short termand long term behaviour. Despite the growing number of field applications,there remain many material and structural implications that need to beaddressed, in particular with regard to long term performance under loads.The procedure for specifying plate and adhesive materials and for obtainingapproval for their use has been simplified in the UK by the amalgamationof the individual product approval systems of the Highways Agency, theBritish Board of Agrément and the County Surveyors’ Society, into one allencompassing scheme known as the Highways Authorities Products Ap-

SOR2 12/19/98, 2:40 PM38


proval Scheme (HAPAS), as described in a recent publication by Roberyand Innes (1997).

Although this book is not specifically concerned with the shear strength-ening of reinforced concrete, some research on this topic has been reportedin this chapter using externally applied composite plates or fabrics; thistechnique also appears to have great potential.

2.8 References

ACI (1973) Use of Epoxy Compounds with Concrete, ACI Committee 503, ACI,Detroit.

Adams R D and Wake W C (1984) Structural Adhesive Joints in Engineering,Elsevier Applied Science, London.

Al-Sulaimani G J, Sharif A, Basunbul I A, Baluch M H and Ghaleb B N (1994)‘Shear repair for reinforced concrete by fiberglass plate bonding’, ACI Struct J91(4) 458–464.

Arduini M, D’Ambrisi Ambrisi A and Di Tommaso A (1994) ‘Shear failure ofconcrete beams reinforced with FRP plates’, in Infrastructure: New Materials andMethods of Repair, ed. K Bashan, Proc 3rd Materials Engineering Conf, ASCE,San Diego, 13–16th Nov 1994, pp 123–130.

Arduini M, Di Tommaso A and Manfroni O (1995) ‘Fracture mechanisms ofconcrete beams bonded with composite plates’, in Non-Metallic (FRP) Reinfor-cement for Concrete Structures, ed. L Taerwe, E & FN Spon, London, pp 483–491.

Baluch M H, Ziraba Y N, Azad A K, Sharif A M, Al-Sulaimani G J and BasunbulI A (1995) ‘Shear strength of plated reinforced concrete beams’, Mag ConcreteRes 47(173) 369–374.

Bencardino F, Spades G and Swamy R N (1997) ‘Design to repair/up-grade RCstructures: the key to a successful utilisation of CFRP laminates’, Proc 7th Inter-national Conference on Structural Faults and Repair, University of Edinburgh,July 8th–10th, 1997, Vol 2, pp 183–190.

Chajes M J, Karbhari V M, Mertz D R, Kaliakin V N and Faqiri A (1993) ‘Rehabili-tation of cracked adjacent concrete box-beam bridges’, Proc Symp PracticalSolutions for Bridge Strengthening and Rehabilitation, NSF, Des Moines, Iowa,1993, pp 265–274.

Chajes M J, Thomson T A, Finch W W and Januszka T F (1994) ‘Flexural streng-thening of concrete beams using externally bonded composite materials’,Construct Build Mater 8(3) 191–201.

Chajes M J, Thomson T A and Tarantino B (1995a) ‘Reinforcement of concretestructures using externally bonded composite materials’, in Non-Metallic (FRP)Reinforcement for Concrete Structures, ed. L Taerwe, E & FN Spon, London,pp 501–508.

Chajes M J, Thomson T A and Farschman C A (1995b) ‘Durability of concretebeams externally reinforced with composite fabrics’, Construct Build Mater 9(3)141–148.

Chajes M J, Finch W W, Januszka T F and Thomson T A (1996) ‘Bond and force

SOR2 12/19/98, 2:40 PM39


transfer of composite material plates bonded to concrete’, ACI Struct J 93(2)208–217.

Char M S, Saadamanesh H and Ehsani M R (1994) ‘Concrete girders externallyprestressed with composite plates’, PCI J (May–June), pp 40, 51.

Deblois M, Picard A and Beaulieu D (1992) ‘Renforcement de poutres en betonarmé à l’aide de matériaux composites: études théoretique et expérimentale’, inAdvanced Composite Materials in Bridges and Structures, eds K W Neale and PLabossière, Proc 1st Int Conf ACMBS, Canadian Society for Civil Engineering,Sherbrooke, Quebec, Oct 1992, pp 265–275.

Deuring M (1993) ‘Verstärken von Stahlbeton mit gespannten Faserver-bundwerkstoffen (Poststrengthening of concrete structures with prestressed ad-vanced composites)’, EMPA Research Report No. 224, EMPA, Dubendorf,CH-8600, Dubendorf, Switzerland (in German).

Deuring M (1994) ‘Brandversuche an nachtraglich verstärken Tragern aus Beton’,EMPA Report No. 148795, EMPA, Dubendorf, CH-8600, Dubendorf, Switzer-land (in German).

Eberline D K, Klaiber F W and Dunker K (1988) ‘Bridge strengthening with epoxy-bonded steel plates’, Transport Res Record (1180) 7–11.

Finch W W, Chajes M J, Mertz D R, Kaliakin V N and Faqiri A (1994) ‘Bridgerehabilitation using composite materials’, in Infrastructure: New Materials andMethods of Repair, Proc 3rd Materials Engineering Conf, ASCE, ed. K BashanSan Diego, 13–16th Nov 1994, pp 1140–1147.

Fleming C J and King G E M (1967) ‘The development of structural adhesives forthree original uses in South Africa’, Proc RILEM Symp Synthetic Resins inBuilding Construction, Paris, Sept 1967, pp 75–92.

Franke L (1986) ‘The dimensioning of adhesive-bonded joints in concrete buildingcomponents’, in Adhesion Between Polymers and Concrete, Proc Int Symp,RILEM, ed. H R Sasse, Aix-en-Provence, Sept 1986, Chapman and Hall, London,pp 461–473.

Garden H N and Hollaway L C (1997) ‘An experimental study of the strengtheningof reinforced concrete beams using prestressed carbon composite plates’, Proc7 th International Conference on Structural Faults and Repair, University ofEdinburgh, July 8th – 10th, 1997, Vol 2, pp 191–199.

Garden H N, Hollaway L C, Thorne A M and Parke G A R (1996) ‘A parameterstudy of the strengthening of reinforced concrete beams with bonded composites’,in Bridge Management 3, eds J E Harding, G A R Parke and M J Ryall, E & FNSpon, London, pp 400–408.

Gilibert Y, Bernasconi J and Colloi C (1976) ‘Mésure des déformations et descontraintes engendrées lors d’un essai de cisaillement en traction à la surface deplaque d’acier collées’, Mater Struct 9(52) 255–265.

Greenfield K T (1995) ‘Carbon fiber laminates for repair of concrete structures’,Mater Perform 34(3) 36–38.

He J H, Pilakoutas K and Waldron P (1997a) ‘CFRP plate strengthening of RCbeams’, Proc 7 th International Conference on Structural Faults and Repair, Univer-sity of Edinburgh, July 8th–10th, 1997a, Vol 2, pp 119–127.

He J H, Pilakoutas K and Waldron P (1997b) ‘Analysis of externally strengthenedR C beams with steel and CFRP plates’, Proc 7 th International Conference onStructural Faults and Repair, University of Edinburgh, July 8th–10th, 1997b, Vol 2,pp 83–92.

SOR2 12/19/98, 2:40 PM40


Hewlett P C and Shaw J D N (1977) in Developments in Adhesives, ed. W C Wake,Applied Science, London, pp 25–75.

Hollaway L C (1993a) Polymer Composites for Civil and Structural Engineering,Blackie Academic and Professional, Glasgow, Scotland.

Hollaway L C (1993b) Use of Advanced Composites in Highway Structures, Reportwritten for the SERC/DTp/LINK Project, Feb 1993.

Hugenschmidt F (1975) ‘Epoxy adhesives for concrete and steel’, in Polymers inConcrete, Proceedings 1st International Congress Polymer Concretes, 7th May1975, The Construction Press, London, 1976, pp 195–209.

Hussain M, Sharif A, Basunbul I A, Baluch M H and Al-Sulaimani G J (1995)‘Flexural behaviour of precracked reinforced concrete beams strengthened exter-nally by steel plates’, ACI Struct J 92(1) 14–22.

Hutchinson A R (1986) Durability of Structural Adhesive Joints, PhD Thesis, Uni-versity of Dundee.

Hutchinson A R and Rahimi H (1993) ‘Behaviour of reinforced concrete beamswith externally bonded fibre reinforced plastics’, Proc 5th International Confer-ence on Structural Faults and Repair, University of Edinburgh, Vol 3, pp 221–228.

Hutchinson A R and Rahimi H (1996) ‘Flexural strengthening of concrete beamswith externally bonded FRP reinforcement’, Proc 2nd International Conferenceon Advanced Composite Materials in Bridges and Structures, Canadian Society forCivil Engineering, Montreal, Quebec, 11–14 Aug 1996.

Hutchinson R, Abdelrahman A and Rizkalla S (1997) ‘Shear strengthening usingCFRP sheets for a prestressed concrete highway bridge in Manitoba, Canada’,Proceedings of the US–Canada–Europe Workshop in Bridge Engineering –Advanced Rehabilitation, Durable Materials, Non-destructive Evaluation andManagement, eds U Meier and R Betti, EMPA, Dubendorf, July 1997, pp 97–104.

Irwin C A K (1975) The Strengthening of Concrete Beams by Bonded Steel Plates,Supplementary Report 160, Transport and Road Research Laboratory,Crowthorne, UK.

Iyer S L (1988) ‘Strengthening of reinforced concrete bridges with external rein-forcements’, Proc 2nd Workshop Bridge Engineering Research in Progress, NSF,Reno, Nevada.

Iyer S L, Sivaramakrishnan C and Atmaram S (1989) ‘Testing of reinforced concretebridges for external reinforcement’, Proc Struct Cong, ASCE, May 1989, pp 116–122.

Jones R and Swamy R N (1995) ‘Plate bonding – What can it do to your beams?’,Construc Repair 9(3) 43–47.

Jones R, Swamy R N and Salman F A R (1985) ‘Structural implications of repairingby epoxy-bonded steel plates’, Proc 2nd International Conference on StructuralFaults and Repair, London, April/May 1985, pp 75–80.

Jones R, Swamy R N and Charif A (1988) ‘Plate separation and anchorage ofreinforced concrete beams strengthened by epoxy-bonded steel plates’, The StructEng 66(5) 85–94.

Kaiser H P (1989) ‘Strengthening of reinforced concrete with epoxy-bonded carbon-fiber plastics’, Doctoral Thesis, Diss. ETH, Nr. 8918, ETH Zurich, Ch-8092 Zu-rich, Switzerland, 1989 (in German).

Karbhari V M and Engineer M (1996) ‘Investigation of bond between concrete andcomposites: use of a peel test’, J Reinforced Plastics Compos 15(2) 208–227.

SOR2 12/19/98, 2:40 PM41


Kinloch A J (1987) Adhesion and Adhesives; Science and Technology, Chapman andHall, London.

Kreigh J D (1976) ‘Arizona slant shear test: a method to determine epoxy bondstrength’, ACI J 73(7) 372–373.

Ladner M and Holtgreve K (1989) ‘Verstärken von Stahlbetonbauten mitaugeklebten CFK-Lamellan’, IABSE Symp, Lisbon, 1989, pp 677–689.

Ladner M and Weder C (1981) Concrete Structures with Bonded External Reinforce-ment, EMPA Report No. 206, EMPA, Dubendorf, CH-8600, Dubendorf,Switzerland.

Ladner M, Pralong J and Weder C (1990) Geklebte Bewehrung: Bemessung undErfahrung, EMPA Report No.116/5, EMPA, Dubendorf, CH-8600, Dubendorf,Switzerland (in German).

Lane J S, Leeming M B, Darby J J and Fashole-Luke P S (1997) ‘Field testing of18 m post tensioned concrete beams strengthened with CFRP plates’, 7 th Int Confon Structural Faults and Repair 97, Edinburgh, July 1997, Vol 1 Extending theLife of Bridges, pp 209–214.

Lee H and Neville K (1967) Handbook of Epoxy Resins, McGraw-Hill, New York.Lees W A (1985) ‘The adhesive bond: a new look at the interaction between

adherend and adhesive’, Plastics Rubber Int 10(3) 11–15.Leibengood L D, Darwin D and Dodds R H (1986) ‘Parameters affecting FE

analysis of concrete structures, J Struc Eng Div ASCE 112 326–341.Lerchenthal C H (1967) ‘Bonded sheet metal reinforcement for concrete slabs’,

Proc RILEM Symp Synthetic Resins in Building Construction, Paris, Sept 1967,pp 165–174.

Lewis A F (1972) ‘Stress endurance limit of lap shear adhesive joints’, AdhesivesAge 15 38–40.

Lewis A F, Kinmouth R A and Kreahling R P (1972) ‘Long term strength ofstructural adhesive joints’, J Adhesion 3(3) 249–257.

L’Hermite R and Bresson J (1967) ‘Beton armé d’armatures collées’, ProcRILEM Symp Synthetic Resins in Building Construction, Paris, Sept 1967, pp 175–203.

Macdonald M D (1978) The Flexural Behaviour of Concrete Beams with BondedExternal Reinforcement, Supplementary Report 415, Transport and Road Re-search Laboratory, Crowthorne, UK.

Macdonald M D (1981) ‘Strength of bonded shear joints subjected to movementduring cure’, Int J Cement Compos Lightweight Concrete 3(4) 267–272.

Macdonald M D (1982) The Flexural Performance of 3.5 m Concrete Beams withVarious Bonded External Reinforcements, Supplementary Report 728, Transportand Road Research Laboratory, Crowthorne, UK.

Macdonald M D and Calder A J J (1982) ‘Bonded steel plating for strengtheningconcrete structures’, Int J Adhesion and Adhesives 2(2) 119–127.

Mackie R I and Su N (1993) ‘The effect of ageing and environment on the static andfatigue strength of adhesive joints’, J Adhesion 42 191–207.

Malvar L J, Warren G E and Inaba C (1995) ‘Rehabilitation of navy pier beams withcomposite sheets’, in Non-Metallic (FRP) Reinforcement for Concrete Structures,ed. L Taerwe, E & FN Spon, London, pp 533–540.

Mays G C (1985) ‘Structural applications of adhesives in civil engineering’, MaterSci Technol 1 937–943.

Mays G C (1993) ‘The use of bonded external reinforcement in bridge strengthen-

SOR2 12/19/98, 2:40 PM42


ing: structural requirements of the adhesive’, in Bridge Management 2, eds JEHarding, GAR Parke and MJ Ryall, Thomas Telford, London, pp 672–680.

Mays G C and Hutchinson A R (1988) ‘Engineering property requirements forstructural adhesives’, Proc ICE 85(2) 485–501.

Mays G C and Hutchinson A R (1992) Adhesives in Civil Engineering, CambridgeUniversity Press, Cambridge.

Meier U (1987) ‘Bridge repair with high performance composite materials’, MaterTechnik 15 125–128 (in French and German).

Meier U (1992) ‘Carbon fiber-reinforced polymers: modern materials in bridgeengineering’, Struct Eng Int 1 7–12.

Meier U (1995) ‘Strengthening of structures using carbon fibre/epoxy composites’,Construct Build Mater 9(6) 341–351.

Meier U and Kaiser H P (1991) ‘Strengthening of structures with CFRP laminates’,Proc Advanced Composite Materials in Civil Engineering Structures, Mats DivASCE, Las Vegas, Jan 1991, pp 224–232.

Meier U and Winistorfer A (1995) ‘Retrofitting of structures through externalbonding of CFRP sheets’, in Non-Metallic (FRP) Reinforcement for ConcreteStructures, ed. L Taerwe, E & FN Spon, London.

Meier U, Deuring M, Meier H and Schwegler G (1992) ‘Strengthening of structureswith advanced composites’, in Alternative Materials for the Reinforcement andPrestressing of Concrete, ed. J L Clarke, Blackie Academic and Professional,Glasgow, pp 153–171.

Moustafa S E (1974) ‘Ultimate load test of a segmentally constructed prestressedconcrete I beam’, J Prestressed Concrete Inst 19(4) 54–75.

Naderi M (1985) Internal Research Report, Civil Engineering Department, Queen’sUniversity, Belfast, Oct 1985.

Naderi M, Clelund D and Long AE (1986) ‘In situ test methods for repairedconcrete structures’, in Adhesion Between Polymers and Concrete, ed. H R Sassa,Proceedings International Symposium RILEM, Aix-en-Provence, Sept 1986,Chapman & Hall, London, pp 707–718.

Nanni A (1995) ‘Concrete repair with externally bonded FRP reinforcement’,Concrete Int 17(6) 22–26.

NCE (1996) ‘Sticking to the task’, New Civil Eng (11 July) 22.Peshkam V and Leeming M (1994) ‘Application of composites to strengthening of

bridges: Project ROBUST’, Proc 19th British Plastics Federation CompositesCong, Birmingham, UK, 22–23 Nov 1994.

Peterson J (1965) ‘Wood beams prestressed with bonded tension elements’, J StructDiv, ASCE 91(1) 103–119.

Priestley M J N, Seible F and Fyfe E (1992) ‘Column seismic retrofit using fibreglass/epoxy jackets’, in Advanced Composite Materials in Bridges and Structures, Proc1st Int Conf ACMBS, eds K W Neale and P Labossière, Canadian Society for CivilEngineering, Sherbrooke, Quebec, Oct 1992, pp 287–298.

Quantrill R J and Hollaway L C (1998) ‘The flexural rehabilitation of reinforcedconcrete beams using prestressed advanced composite plates’, Composite Scienceand Technology 58(8).

Quantrill R J, Hollaway L C, Thorne A M and Parke G A R (1995) ‘Preliminaryresearch on the strengthening of reinforced concrete beams using FRP’, in Non-Metallic (FRP) Reinforcement for Concrete Structures, ed. L Taerwe, E & FNSpon, London, pp 543–550.

SOR2 12/19/98, 2:40 PM43


Quantrill R J, Hollaway L C and Thorne A M (1996a) ‘Part 1. Experimental andanalytical investigation of FRP strengthened beam response’, Mag Concrete Res48(177) 331–342.

Quantrill R J, Hollaway L C and Thorne A M (1996b) ‘Part 2. Predictions of themaximum plate end stresses of FRP strengthened beams’, Mag Concrete Res48(177) 343–352.

Raithby K D (1980) External Strengthening of Concrete Bridges with Bonded SteelPlates, Supplementary Report 612, Transport and Road Research Laboratory,Crowthorne, UK.

Raithby K D (1982) ‘Strengthening of concrete bridge decks with epoxy-bondedsteel plates’, Int J Adhesion Adhesives 2(2) 115–118.

Roberts T M (1989) ‘Approximate analysis of shear and normal stress concentra-tions in the adhesive layer of plated RC beams’, The Struct Eng 67(12) 229–233.

Robery P and Innes C (1997) ‘Carbon fibre strengthening of concrete structures’,Proc 7 th International Conference on Structural Faults and Repair, University ofEdinburgh, July 8th–10th 1997, Vol 1, pp 197–208.

Rostasy F S, Hankers C and Ranisch E H (1992) ‘Strengthening of R/C- and P/C-structures with bonded FRP plates’, in Advanced Composite Materials in Bridgesand Structures, eds K W Neale and P Labossière, Proc 1st Int Conf ACMBS,Canadian Society for Civil Engineering, Sherbrooke, Quebec, Oct 1992, pp 253–263.

Saadatmanesh H and Ehsani M R (1989) ‘Applications of fiber-composites in civilengineering’, Proc Sessions Related to Structural Materials, Structures Cong,ASCE, 1989, pp 526–535.

Saadatmanesh H and Ehsani M R (1990a) ‘Fiber composite plates can strengthenconcrete beams’, Concrete Int ACI 12(3) 65–71.

Saadatmanesh H and Ehsani M R (1990b) ‘Flexural strength of externally rein-forced concrete beams’, Proc 1st Materials Engineering Cong, ASCE, 1990, pp1152–1161.

Saadatmanesh H and Ehsani M R (1991) ‘RC beams strengthened with GFRPplates I: experimental study’, J Struct Eng 117(11) 3417–3433.

Santoh N, Kimura H, Enomoto T, Kiuchi T and Kuzubi Y (1983) ‘Report on theuse of CFCC in prestressed concrete bridges in Japan’, Proc Int Symp Fiber-Reinforced-Plastic Reinforcement for Concrete Structures, eds A Nanni and C WDolan, ACI SP-138, ACI, Detroit, pp 895–911.

Sasse H R and Friebrich M (1983) ‘Bonding of polymer materials to concrete’,Mater Struct 16(94) 293–301.

Schutz R J (1976) ‘Epoxy adhesives in prestressed and precast concrete bridgeconstruction’, ACI J 155–159.

Shaw J D N (1982) ‘A review of resins used in construction’, Int J AdhesionAdhesives 2(2) 77–83.

Shaw M (1993) ‘Strengthening bridges with externally bonded reinforcement’, inBridge Management 2, eds J E Harding, G A R Parke and M J Ryall, ThomasTelford, London, pp 651–659.

Shields J (1985) Adhesives Handbook, 3rd edn, Newnes-Butterworth, London.Solomon S K (1976) A Study of Steel and Concrete Sandwich Construction for Use

in Bridge Roadway Decks, PhD Thesis, University of Dundee, Jan 1976.Swamy R N and Mukhopadhyaya P (1995) ‘Role and effectiveness of non-metallic

plates in strengthening and upgrading concrete structures’, in Non-Metallic (FRP)

SOR2 12/19/98, 2:40 PM44


Reinforcement for Concrete Structures, ed. L Taerwe, E & FN Spon, London, pp473–481.

Tabor L J (1978) ‘Effective use of epoxy and polyester resins’, in Civil EngineeringStructures, CIRIA Report 69, Construction Industry and Information Association,London, Jan 1978.

Tabor L J (1982) ‘Adhesives in construction’, Int J Adhesion Adhesives, April 1982,73–76.

Tabor L J (1985) ‘Twixt old and new: achieving a bond when casting fresh concreteagainst hardened concrete’, Proc 2nd Int Conf Structural Faults and Repair,Institution of Civil Engineers, London, April/May, 1985, pp 57–63.

Taljsten B (1997) ‘Strengthening of concrete structures for shear with bondedCFRP-fabrics’, Proceedings of the US-Canada-Europe Workshop on RecentAdvances in Bridge Engineering – Advanced Rehabilitation, Durable Materials,Non-destructive Evaluation and Management, eds U Meier and R Betti, EMPA,Dubendorf, July 1997.

Triantafillou T C and Deskovic N (1991) ‘Innovative prestressing with FRP sheets;mechanics of short term behaviour’, J Eng Mech 117(7) 1652–1671.

Triantafillou T C and Plevris N (1992) ‘Strengthening of RC beams with epoxy-bonded fibre-composite materials’, Mater Struct 25 201–211.

Triantafillou T C and Plevris N (1995) ‘Reliability analysis of reinforced concretebeams strengthened with CFRP laminates’, in Non-Metallic (FRP) Reinforcementfor Concrete Structures, ed. L Taerwe, E & FN Spon, London, pp 576–583.

Triantafillou T C, Deskovic N and Deuring M (1992) ‘Strengthening of concretestructures with prestressed fiber reinforced plastic sheets’, ACI Struct J 89(3)222–235.

Tu L and Kruger D (1996) ‘Engineering properties of epoxy resins used as concreteadhesives’, ACI Mater J 93(1) 26–35.

Varastehpour H and Hamelin P (1995) ‘Structural behaviour of reinforced concretebeams strengthened by epoxy bonded FRP plates’, in Non-Metallic (FRP) Rein-forcement for Concrete Structures, ed. L Taerwe, E & FN Spon, London, pp 559–567.

Vardy A E and Hutchinson A R (1986) ‘Some uses of adhesives in civil engineer-ing’, Proceedings, International Conference on Structural Adhesives in Engineer-ing, Institution of Mechanical Engineers, Bristol University, July 1986, pp199–204.

Wake W C (1982) Adhesion and the Formulation of Adhesives, 2nd edn, AppliedScience Publishers, London.

Wall J S, Shrive N G and Gamble B R (1986) ‘Testing of bond between fresh andhardened concrete’, in Adhesion Between Polymers and Concrete, Proc Int Symp,RILEM, Aix-en-Provence, Sept 1986, ed. H R Sasse, Chapman and Hall, London,pp 335–344.

Wight R G, Green M F and Erki M A (1995) ‘Post-strengthening concrete beamswith prestressed FRP sheets’, in Non-Metallic (FRP) Reinforcement for ConcreteStructures, ed. L Taerwe, E & FN Spon, London, pp 568–575.

SOR2 12/19/98, 2:40 PM45


46

3Materials

A R H U T C H I N S O N A N D J Q U I N N

3.1 Adhesive bonded connections

The primary purpose of a connection is to transfer load between twocomponents. The ability to join components comprising dissimilar materialsefficiently using adhesives represents one of the key motivations for thewhole concept of strengthening using the plate bonding technique. The mostcommonly used generic type of adhesive used for this purpose is an epoxy.

3.1.1 General considerations

The basic requirements for the creation of a satisfactory bonded joint are:

• selection of a suitable adhesive• adequate preparation of the adherend surfaces• appropriate design of the joint• controlled fabrication of the joint itself.

Some form of postbonding quality assurance is also desirable.Conceptually, adhesives represent natural candidates for joining FRP

materials. This is because the resin matrix which is used as a binder incomposite materials is itself an organic adhesive, and adhesives which arethen used to join composite materials together may themselves be verysimilar in terms of their chemical composition and mechanical properties.For instance, ‘like’ is used to join ‘like’ when bonding an epoxy matrixcomposite with an epoxy adhesive. The main difference here is the lowercuring temperature (say 15 °C) used to cure the adhesive, compared to therather higher temperature (say 120–180 °C) used to cure the fibre rein-forced vinyl ester or epoxy matrix resin.

There are two particular problems associated with adhesive bonding ofFRP materials:

• attachment to the surface of a layered material• the surface may be contaminated with mould release agents.

SOR3 12/19/98, 2:43 PM46

Materials 47

Figure 3.1 Bondline strain variations in a single cover butt shear jointunder load

To these must be added the problems associated with adhesive bondingof concrete:

• attachment to the surface of a relatively weak and brittle material• the surface may be cracked• the surface may be cement rich and contaminated• concrete is highly alkaline and may have a high moisture content.

Wherever possible, joint designs should limit the problems associatedwith low tensile strength of the concrete, low through-thickness strengthand potentially weak surface layers of fibre reinforced plastic (FRP) ma-terials. Thus careful joint design is required to minimise cleavage forcesacting on bonded configurations and bond interfaces. Adequate surfacepreparation is also vital and contaminants such as mould release agents anddust must be removed from all surfaces prior to the application of a suitableadhesive.

3.1.2 Characteristics of bonded joints

The transfer of load between two components through an adhesive layerresults in bondline stress (strain) variations. This is illustrated in Fig. 3.1 fora single cover butt shear joint. It is apparent that the adhesive and theinterfaces are loaded in shear. Such a joint also tends to peel apart as shown,because the line of action of the loads is not coincident. Peel forces can bevery high and bonded joint design should embrace those geometries whichminimise the generation of peel forces.

3.2 Composite materials

3.2.1 Composition

The primary materials in a composite are the reinforcement fibre and thepolymer matrix. Other materials are incorporated in the composite but theyare of less significance in terms of both effect on cost and effect on proper-ties. Although the term polymer composites includes both thermosetting

SOR3 12/19/98, 2:43 PM47


and thermoplastic resins, we are only concerned with the former in thisbook.

The most commonly used thermosetting resins in composites are polyes-ter, urethane methacrylate, vinylester, epoxy and phenolic. They are iso-tropic materials which allow load transfer between the fibres, but theyperform several other duties. The matrix protects notch-sensitive fibresfrom abrasion and it forms a protective barrier between the fibres and theenvironment, thus preventing attack from moisture, chemicals and oxida-tion. It also plays an important role in providing shear, transverse tensileand compression properties. The thermomechanical performance of thecomposite is also governed by the matrix performance.

Reinforcement fibres fall into three main families of glass, aramid andcarbon. There are other fibres but they are relatively insignificant. The mostimportant property of the fibres is their elastic modulus, and the fibres mustbe significantly stiffer than the matrix which allows them to carry most ofthe stress. Consequently they must also be of high strength. Reinforcementsare available in a variety of configurations of which there are three maincategories:

• unidirectional, in which all the fibres lie in one direction.• bidirectional, in which the fibres lie at 90° to one another. This is

achieved either by the use of woven fabric, non-woven fabric or by theuse of separate layers of fibres each unidirectional but successively laidat 90°.

• random, in which the fibres are randomly distributed and are in-plane.

The ROBUST Project used unidirectional reinforcement.

3.2.1.1 Glass fibres

As a broad generalisation, glass fibres can be categorised into two sets.Those with a modulus around 70GPa and with strengths after processing inthe range 1000–2000MPa (i.e. E, A, C, E-CR), and those with a modulusaround 85GPa with strengths after processing in the range 2000–3000 MPa(i.e. R, S and AR). These fibre types and their principal uses are shown inTable 3.1.

The density of glass fibre is about 2500kg m�3, which is high in compari-son to other reinforcing fibres but, by metallic standards, is very low (alu-minium has a density of about 2800kgm�3 and steel 7800kgm�3). Glassfibres are the most commonly used reinforcing fibres and there are severalvery good reasons for this. They have good properties both in an absolutesense and relative to weight. They also have very good processing charac-teristics and they are inexpensive.

SOR3 12/19/98, 2:43 PM48

Materials 49

Table 3.1 Types of glass fibre and principal uses

Fibre type Principal use

E Standard reinforcement, low alkali content (�1%)A High alkali content (10–15%) but inferior properties to type EC Superior corrosion resistance to type E, often used in the

form of a surface veil or tissueE-CR Boron-free, good acid corrosion resistance, with similar

properties to type ER, S Better mechanical properties than type E, used for higher

performance applicationsAR Alkali-resistant, used for reinforcement of cement

The processing characteristics of particular types of glass fibre have beenmodified and optimised over many years to achieve the required perform-ance, such as choppability, low static buildup and conformance to complexshape, and resin compatibility requirements such as fast wet out, good fibre/matrix adhesion, and so on.

Glass fibres and composites made with them are to a certain extent proneto moisture-induced degradation, particularly at significant sustained strainlevels. Surface corrosion of fibres and the resultant reduction in fibre/matrixadhesion is known as stress-corrosion cracking.

3.2.1.2 Carbon fibres

Carbon fibres are the predominant reinforcement used to achieve highstiffness and high strength. The term carbon fibre (graphite fibres in theUSA) covers a whole family of materials which encompass a large rangeof strengths and stiffness. The density of carbon fibre is of the order of1900 kgm�3. Typical fibre moduli may be 230–300GPa, whilst strengthsafter processing are in the range of 3000–5000 MPa.

Carbon fibre is most commonly produced from a precursor ofpolyacrylonitrile (PAN) fibre which is processed by first stretching it toachieve a high degree of molecular orientation. It is then stabilised in anoxidising atmosphere while held under tension. The fibres are then sub-jected to a carbonising regime at a temperature in the range 1000–3500 °C.The degree of carbonisation determines such properties as elastic modulus,density and electrical conductivity. As an alternative to the use of PAN,routes via the use of pitch and rayon have been successfully utilised andsuch fibres are commercially available. These fibres tend to be of lowerperformance than PAN-based fibres; they are also cheaper due to their useof a lower cost precursor.

SOR3 12/19/98, 2:43 PM49


3.2.1.3 Aramid fibres

Aramid fibres are very tough organic synthetic fibres generally character-ised as having reasonably high strengths up to 3000MPa, moduli in therange 60–120GPa, and a very low density (around 1400 kgm�3). Compositesmade with aramid fibres fit well into a gap in the range of stress/straincurves left by the family of carbon fibres at one extreme and glass fibres atthe other.

Aramid fibres are fire resistant and perform well at high temperatures.They are insulators of both electricity and heat and are resistant to organicsolvents, fuels and lubricants. A major distinction of aramid fibres is thatthey are highly tenacious in the non-composite form and do not behave ina brittle manner as do both carbon and glass fibres.

The tensile stress/strain curve of aramid fibres is essentially linear tofailure. They have two distinct categories; those in which their elastic modu-lus is about the same as glass fibre, typically 60–70GPa, and those witha modulus at about twice this level. Kevlar 29 falls into the first categoryand Kevlar 49 into the higher modulus category. It is generally the highermodulus material which finds use in composites, but the lower modulusaramids do have applications in composites in those circumstances wherehigh strain-to-failure or high work-to-failure are required. The specificperformance of aramids is their primary advantage, that is their strength/weight and stiffness/weight ratios. The density of aramids exhibits advan-tage over many carbon and glass fibres if either specific strength or specificstiffness is the selection criterion. Some aramids have relatively very lowcompressive strength.

3.2.2 Manufacturing routes

The manufacturing options available are diverse and have been developedto suit the wide variety of production parameters encountered. They maybe classified as shown in Table 3.2.

A full discussion of these options and typical properties of compositesmanufactured using these processes is given by Hollaway (1993). The pro-cesses used in the ROBUST Project were autoclave/vacuum bag andpultrusion.

3.2.2.1 Vacuum bag/autoclave moulding

Layers of prepreg (resin impregnated fibre reinforcement) are applied to amould and rolled. A rubber or nylon sheet or bag is placed over the lay-upand the air is removed by means of a vacuum pump. The mould is thenplaced in an oven where heat (typically 80–200 °C) is applied, or into anautoclave which applies both heat and pressure (typically 7bar).

SOR3 12/19/98, 2:43 PM50

Materials 51

Table 3.2 Classification of composite manufacturing options

Class Process

Open lay-up Hand lay-upSpray up

Intermediate Cold pressResin transfer moulding (RTM)/resin injectionAutoclave/vacuum bag

Compression Hot press

Continuous PultrusionContinuous sheet

Winding processes Filament winding

Figure 3.2 The pultrusion process

As an alternative to prepreg, dry reinforcement may be used which isthen impregnated with resin on the mould. A cold cure resin system may beused with this method, that is the resin mix contains an accelerator as wellas a catalyst. This is a useful option if a heat source is not available. Theprocess is slow and laborious but produces excellent properties.

3.2.2.2 Pultrusion

Pultrusion is a continuous process which allows the production of straight,constant section profiles. Reinforcement is pulled through a device (typi-cally a bath) which applies the resin. The wetted fibre is then pulled onthrough a heated steel die which is the shape of the section to be produced.The die is heated to about 150 °C which causes the resin to react, gel andcure. When the profile exits from the die it has already achieved a highdegree of cure. The profile is pulled by either reciprocating pullers or acaterpillar haul-off and it is then automatically sawn to length. The processis illustrated schematically in Fig. 3.2. Resin systems must be highly reactiveto cure in the available time. Machine speed, die temperature and resinreactivity are parameters which interact and must be balanced. Releaseagents are commonly dissolved in the resin and these migrate to the surface

SOR3 12/19/98, 2:43 PM51


during cure in order to prevent the artefact from sticking to the surface ofthe forming dies. As the process is ‘hot cure’, an accelerator is not required.

Most of the reinforcement configurations may be used but in generalpultrusions are dominated by the use of unidirectional reinforcement,which lends itself most appropriately to the process and gives maximumstrength and stiffness in the axial direction of the profile. Fibre volumesof up to 65% are achievable with unidirectionally aligned fibres. A peel-plyfabric may be incorporated on the outermost surface(s) to provide bothprotection and a sacrificial surface treatment (see Section 3.4.5).

3.2.3 Manufacturing quality

It is sometimes argued that composites tend to be variable in nature due tothe manner in which they are manufactured. This is a broad generalisationwhich does not apply to several composite types including prepreg mould-ings and pultrusions, particularly those in which the reinforcement is uni-directional. These are the types of material used in the ROBUST Projectand these materials can be expected to be highly consistent.

The performance of joints made with composite materials bonded in situis dependent upon the quality of the supplied materials, both composite andadhesive, and the manner in which bonding is carried out. The latter is morelikely to generate problems due to such factors as variations in temperatureand humidity, and the likelihood of occurrence of dust, moisture, oils, andso on.

It is essential in the adhesive bonding process that a ‘method of work’ isused. This must take into account the possible working conditions whichmay occur and state the required procedure for each set of conditions. Itmust also state the conditions under which adhesive bonding may not takeplace.

3.2.4 Typical properties

Composite materials are not homogeneous. Their properties are dependenton many factors, the most important of which are the type of fibre, quantityof fibre (as volume fraction) and the configuration of the reinforcement.They are generally completely elastic up to failure and exhibit neither ayield point nor a region of plasticity. They tend to have low strain to failure(less than 3%). The resulting area under the stress/strain curve, whichrepresents the work done to failure, is relatively small when compared tomany metals.

The properties of composites are dependent on the properties of boththe fibre and the matrix, the proportion of each and the configuration ofthe fibres. If all the fibres are aligned in one direction then the composite is

SOR3 12/19/98, 2:43 PM52

Materials 53

Figure 3.3 Theoretical variation in tensile modulus with the angleof load relative to the principal fibre direction (unidirectionalcarbon fibre reinforced plastic (UD CFRP), fibre volume fractionVf � 0.5).

relatively stiff and strong in that direction, but in the transverse directionit has low modulus and low strength. When a unidirectional composite istested at a small angle from the fibre axis, there is a considerable reductionin strength. A similar but less significant effect occurs with the tensilemodulus, as shown in Fig. 3.3.

3.2.4.1 Strength and stiffness

3.2.4.1.1 Glass fibre reinforced polymer (GFRP)

E glass fibres, which have a modulus of about 70GPa, produce compositeswith modest moduli. In the case of unidirectional fibres and the highesttypical fibre volume fraction of 0.65, a composite has a modulus of about45 GPa and a strength of around 1300 MPa. At right angles to this, in thetransverse direction, the modulus approaches that of the resin itself at about4 GPa and the strength would be 50–100MPa. The unidirectional compo-sites used in the ROBUST Project, manufactured using the vacuum bag

SOR3 12/19/98, 2:43 PM53


process with prepreg material in an epoxy matrix, had the followingproperties:

• longitudinal tensile modulus: 36GPa• longitudinal tensile strength: 750 MPa• elongation at break: 3.1%.

Bidirectional E glass laminates have a typical fibre volume fraction ofabout 0.4 and a tensile modulus at that volume fraction of about 14GPa.Random laminates (e.g. chopped strand mat) have a typical fibre volumefraction of about 0.2 and a tensile modulus at that volume fraction of about9GPa. The use of S2 or R glass improves the composite modulus to about60GPa for unidirectional and 20 GPa for woven fabric (bidirectional) con-structions. This is at some monetary disadvantage. They are both moreexpensive than E glass and they are only available in a fairly limited rangeof material types and resin compatibilities. Probably the most importantvirtue of S2 and R glass is their high strength, which is considerably higherthan E glass.

3.2.4.1.2 Carbon fibre reinforced plastic (CFRP)

The dominant carbon fibres in current use (typically Toray T700) have atensile modulus of about 230GPa, a tensile strength of around 5000MPaand a strain-to-failure of 2%. Unidirectional composites produced fromthem in either an epoxy or vinylester matrix have the following typicalproperties:

• longitudinal tensile modulus: 155–165GPa• longitudinal tensile strength: 2500–3000 MPa• elongation at break: 1.2–1.3%

Carbon fibres are available which will give a tensile modulus of about250GPa in a unidirectional composite, comparing very favourably withsteel at about 210GPa. However, as this composite is unidirectional, it hasextremely low modulus in the transverse direction. The principal attributesof carbon fibre composites are their very high specific stiffness (the ratio ofmodulus/density), excellent fatigue and environmental resistance.

Currently there are various pultruded CFRP plates available commer-cially for plate bonding applications. The pultruded plates used in theROBUST Project, as well as the plates manufactured with prepreg ma-terials, possessed a modulus of about 130GPa and a strength of 1500MPa.Pultruded plates now available from other sources typically exploit fibreswith superior properties such as Toray T700, resulting in composites withthe properties shown above. A financial penalty has to be paid for materials

SOR3 12/19/98, 2:43 PM54

Materials 55

exhibiting properties significantly in excess of these; the strain-to-failure ofcomposites made with them will also be reduced significantly.

3.2.4.1.3 Aramid fibre reinforced plastic (AFRP)

Unidirectional aramid composites have high tensile strength (1200–1400 MPa) and a very low density. This high specific tensile strength is animportant attribute which makes them particularly suited to use as tensionmembers. Some aramids exhibit a relatively low compressive yield strengthof about 230MPa. Thus composites using these fibres must be carefullydesigned, particularly for compression or bending. This makes themparticularly suited to use in tension member applications, but often notsuited to bending applications. It is usually the high modulus variants ofaramids (e.g. Kevlar 49) which are most commonly used as compositereinforcement, conferring a tensile modulus for unidirectional compositesof about 75 GPa. This is very similar to aluminium; however, as this is aunidirectional composite the associated transverse modulus is only about5 GPa.

The tensile modulus of unidirectional and bidirectional aramid compo-sites represents a reasonable compromise between the low modulus glassfibre composites and the much higher carbon fibre composites. Similarly thetensile strength of aramid composites is a compromise between E glass andcarbon fibre composites.

3.2.4.2 Fatigue

E glass composites have a relatively low modulus which often results inthem experiencing strains which approach the cracking strain of the matrix.This allows a fatigue process to occur which may result in a reduced fatiguelife. It is therefore prudent to use a resin system with high strain-to-failure.Therefore as a general rule isophthalic resin is better than the orthophthalicresin and vinylester resin is significantly better than both.

The fatigue performance of carbon fibre composites is generally far su-perior to both metals and other composites. This is particularly the casefor unidirectional composites loaded in the fibre direction, where they arerelatively insensitive to tension fatigue damage even at very high stresslevels. This does not necessarily apply to all constructions. The fatigue lifeis reduced if the performance of the matrix determines performance suchas �45° lay-up composites or when the stress is transverse to the fibredirection.

The fatigue behaviour of aramids can be excellent, but under othercircumstances is very poor. For instance, unidirectional Kevlar 49composites in tension/tension fatigue are superior to S glass and E glass

SOR3 12/19/98, 2:43 PM55


composites and 2024-T3 aluminium; only unidirectional carbon compositesare superior. However in flexural fatigue, at a fairly low number of cycles,they have poorer resistance than E glass due to the innate poor staticflexural performance of Kevlar 49.

3.2.4.3 Creep and stress rupture

Unidirectional composites subjected to tensile load resist long term creepvery well. The total elastic performance of the fibres dominates the behav-iour of the composite, but, when under compression or ‘off-axis’, then thematrix properties take on a more significant role. Polymers are viscoelasticmaterials and they deflect continuously with time under load. Under thesecircumstances, and particularly at elevated temperature, the creep perform-ance of the composite requires particular consideration.

The most creep-resistant construction is unidirectional reinforcement,followed by 0,90° non-woven construction. Woven construction of 0,90°fibres has the disadvantage that the fibres tend to straighten out, thusincreasing creep. Laminates with a random fibre array are the least resistantto creep, but this may be alleviated to some extent by the use of a highvolume fraction of fibre.

The creep performance of carbon fibre composites in the fibre directionis comparable with ‘low relaxation’ steel and is significantly better than‘standard steel’. In spite of their high inherent tensile strength, and even inunidirectional configurations, aramid fibre composites exhibit creep ratesgenerally very much higher than glass or carbon composites.

3.2.4.4 Thermal expansion coefficients

The thermal expansion of composites is dependent on several factors; thetype of reinforcement, type of matrix, geometry of reinforcement andvolume fraction. Table 3.3 shows the thermal expansion coefficients for avariety of composites. The two negative values are a result of the negativefibre expansion coefficients which are exhibited by carbon and aramid

Table 3.3 Typical thermal expansion coefficients (�10�6 °C)

Fibre/matrix 0° 90° 0,90°

E glass/polyester 8.6 14 9.8Carbon T300/epoxy �0.3 28 1.9Aramid K49/epoxy �4 79 3.2

Note: volume fraction of fibres � 0.65.

SOR3 12/19/98, 2:43 PM56

Materials 57

fibres. It is possible to tailor the expansion coefficient of a composite to aspecific requirement, for example to match that of aluminium, or even to bezero over a given temperature range.

3.2.5 Handling and machining

Composite plates used for externally bonded reinforcement have a highproportion of the fibre aligned with the axis of the plate. This implies highstrength and stiffness in that direction but, in the transverse direction(across the plate), it will be relatively weak and prone to cracking if mishan-dled. Thus care must be taken, for example when cutting to length to ensurethat longitudinal cracks are not started. Although composite materials maybe machined with hand tools such as pistol drills and hacksaws, they mayproduce cracking in composites which have a large percentage of unidirec-tional fibres. Sawing to length, for instance, is often better achieved usingdiamond-tipped blades which have a less destructive tendency than, say, ahacksaw. Further details are given by Hutchinson (1997).

Composite reinforcing plates are sufficiently flexible to allow them to becoiled to a diameter of 1–2 m which makes them easy to transport. Howeverthe coiling process stores a significant amount of energy which could bedangerous if released in an uncontrolled manner. Care must therefore betaken when removing the coiling constraints.

3.3 Adhesive materials

3.3.1 General considerations

The purpose of the adhesive is to produce a continuous bond between fibrereinforced polymer (FRP) and concrete to ensure that full composite actionis developed by the transfer of shear stress across the thickness of theadhesive layer. For this to occur, an excellent degree of adhesion to thesurfaces involved must be achieved and sustained. Experience has shownthat the best chance of success is likely to be achieved by using two-partcold curing paste epoxy adhesives which have been specially developed foruse in the construction industry. Mays and Hutchinson (1988, 1992) identi-fied the principal requirements for bonding steel to concrete and theseare very similar for the case of bonding FRP to concrete. In summary, theadhesive requirements are that:

• It should exhibit adequate adhesion to the materials involved (in thiscase, concrete and FRP).

• A two-part epoxy resin with a polyamine-based hardener should beused, which exhibits good moisture resistance and resistance to creep.

SOR3 12/19/98, 2:43 PM57


• The glass transition temperature of the adhesive, Tg, should be at least40 °C.

• The flexural modulus of the material should fall within the range of2–10 GPa at 20 °C. Generally held views are that the lower boundary ofthis range might be reduced to 1GPa.

• The bulk shear and tensile strength at 20 °C should be � 12MPa.• The minimum shear strength at 20 °C, measured by the thick adherend

shear test (TAST), should be 18 MPa.• The mode I fracture toughness (K1c) should be � 0.5MN m�3/2.• The equilibrium water content (M∝ ) should not exceed 3% by weight

after immersion in distilled water at 20 °C. The coefficient of permeabil-ity should not exceed 5 � 10�14 m2 s�1.

• It should possess gap-filling properties, be thixotropic, and be suitablefor application to vertical and overhead surfaces.

• It should not be sensitive to the alkaline nature of concrete and itspotential effect on the durability of joints.

• It should not be unduly sensitive to variations in the quality or moisturecontent of prepared surfaces.

• The working characteristics of the material should enable an adequatejoint quality to be achieved with respect to mixing, application andcuring.

A further item to add to this list is that the adhesive should exhibitsufficient tack or ‘grab’ to enable FRP materials to be attached to overheador vertical surfaces without the need for temporary fixings whilst the adhe-sive cures.

The mechanical and thermal properties of an adhesive should also beconsidered in relation to those of concrete and FRP materials. The effectsof environmental and other service conditions on the adhesive material andon the behaviour of bonded joints, must also be considered carefully (seeSection 3.3.2 and Chapter 6).

The detailed effect of the mechanical properties of the adhesive on theoverall performance of externally strengthened beams is not yet fully un-derstood. Whilst bonding systems have been, and continue to be, usedsuccessfully, the optimum properties have not been identified. For example,relatively ductile adhesives are usually preferred for joining FRP materials(e.g. Hutchinson, 1997) in order to minimise surface stress concentrations.However, to date high modulus untoughened construction epoxies havebeen used for plate bonding which are characterised by their relatively lowductility. Intuitively, more flexibilised products would be preferred whichpossess superior toughness and increased strain to failure – provided thattheir other desirable properties are not compromised.

SOR3 12/19/98, 2:43 PM58

Materials 59

3.3.2 Epoxy adhesives

Epoxies represent an important class of thermosetting adhesives whichhave been available commercially since the 1940s. As structural adhesives,they are the most widely used and accepted. Other thermosetting adhesivesinclude polyesters and particular variants of acrylics and polyurethanes. Adetailed consideration of adhesives and their properties is given by Maysand Hutchinson (1992).

Epoxy adhesives can be formulated in a variety of forms to provide abroad range of application characteristics and mechanical propertieswhen cured. Commercial formulations are, in general, complex and aresophisticated blends of many components, the most important being theresin. One of a range of different types of hardener is added to thebase resin, together with fillers, toughening agents, plasticizers, diluents,surfactants, antioxidants and other materials. The choice of hardener re-sults in distinction between the adhesives which cure either at ambient orat elevated temperature. However as a rule of thumb, the rate of chemicalreaction is doubled for every 8 °C rise in temperature. The blending of thebase resin with a variety of materials indicates that the possibilities foraltering the properties and characteristics of an epoxy adhesive are numer-ous. On the negative side it should be said that blending with a variety ofadditives tends to make any adhesive more expensive than its unmodifiedcounterpart.

Epoxy resins have several advantages over other polymers as adhesiveagents for civil engineering use, namely:

• High surface activity and good wetting properties for a variety ofsubstrates.

• May be formulated to have a long open time (the time between applica-tion and closing of the joint).

• High cured cohesive strength; joint failure may be dictated by adherendstrength (particularly with concrete substrates).

• May be toughened by the inclusion of a dispersed rubbery phase.• Lack of by-products from curing reaction minimises shrinkage and al-

lows the bonding of large areas with only contact pressure.• Low shrinkage compared with polyesters, acrylics and vinyl types;

hence, residual bondline strain in cured joints is reduced.• Low creep and superior strength retention under sustained load.• Can be made thixotropic for application to vertical surfaces.• Able to accommodate irregular or thick bondlines (e.g. concrete

adherends).• May be modified by (a) selection of base resin and hardener; (b) addition

of other polymers; (c) addition of surfactants, fillers and other modifiers.

SOR3 12/19/98, 2:43 PM59


Typically they are suitable for service operating temperatures in the range�60 °C to �60 °C.

The properties of Sikadur 31PBA, the filled epoxy adhesive used in theROBUST system of CFRP plate bonding, are discussed in Section 3.3.5 andin Chapter 11.

3.3.2.1 Resins

Epoxy resins are traditionally based upon the reaction of epichlorohydrinon bis-phenol A, to give a liquid compound of linear molecules terminatingwith epoxy groups and having secondary hydroxyl groups occurring atregular intervals. The resulting base resin is known as the diglycidyl ether ofbis-phenol A (DGEBA) and is commonly encountered in two-part roomtemperature cure epoxy adhesives used in construction. The adhesive prop-erties of epoxies are obtained by polymerisation using a cross-linking agent,or hardener, to form a tough three-dimensional polymer network.

3.3.2.2 Hardeners

Aliphatic polyamines are one of the most commonly used hardeners forambient temperature curing epoxy adhesives, conferring a rigid structurewith good resistance to chemicals, solvents and water. A common modifi-cation which is performed commercially to improve the performance selec-tively is to react a glycidyl ether resin with an excess of amine groups toproduce a resinous amine adduct hardener. These have advantages overunmodified versions in that they result in more convenient mixing ratios,are less hazardous to handle and can exhibit reduced moisture sensitivity.Sikadur 31PBA, the epoxy adhesive used as part of the ROBUST system ofCFRP plate bonding, is cured with an aliphatic polyamine adduct.

3.3.2.3 Epoxy additives

• Fillers: in practice most epoxy resin systems have fillers incorporated,often simply to reduce cost although they may also assist in gap-filling,reduction of creep, reduction of exotherm, corrosion inhibition and fireretardation. They will also alter the physical and mechanical propertiesof the adhesive. Construction resins in particular often include a largevolume fraction of sand or silica.

• Diluents: these materials reduce viscosity of the freshly mixed adhesiveto offset the effect of the filler. Pot life, flexibility and even the Tg of thecured adhesive can also be altered.

• Flexibilisers: these may be used to improve impact resistance, peelstrength and ductility.

SOR3 12/19/98, 2:43 PM60

Materials 61

• Tougheners: unmodified epoxies tend to be relatively brittle whencleavage or peel forces are imposed. Different types of tougheningagents can be added to the polymer structure to improve fracture en-ergy, G1 and therefore the fracture toughness, K1.

• Adhesion promoters: these are sometimes incorporated, both to in-crease the filler–polymer adhesion and to enhance resin adhesion tosurfaces such as glass or metals.

3.3.3 Primers and coupling agents

With some bonding surfaces the application of an adhesive-compatibleprimer coating may be desirable. In steel plate bonding applications it isvirtually essential to prime the steel in order to generate a reliable andreproducible surface which confers good bond stability to an adhesive in thepresence of moisture. No such primers are usually necessary for bondingFRP materials. Conceptually it might be desirable to prime cracked orporous surfaces such as concrete to generate suitable conditions for theapplication of a relatively viscous adhesive. In fact, the use of primers forconcrete is rarely reported in structural adhesive bonding unless it is sus-pected that the alkaline nature of the substrate will result in degradationof the adhesive. One of the advantages of using a well-formulated epoxyadhesive is that it should be tolerant of such alkalinity, and will in fact sealand adhere to a concrete surface satisfactorily.

3.3.4 Properties of adhesives

The engineer will be concerned about the behaviour and performance ofthe selected adhesive from the time it is purchased from the manufacturer,through the mixing, application and curing phases to its properties in thehardened state within a joint over the intended design life. Thus the prop-erties of interest in approximate chronological order are likely to include:

• unmixed – shelf life• freshly mixed – viscosity, usable life, wetting ability, joint open time• during cure – rate of strength development• hardened – strength and stress/strain characteristics, fracture toughness,

temperature resistance, moisture resistance, creep, fatigue.

3.3.4.1 Shelf life

This is the period for which the unmixed components may be stored withoutundergoing significant deterioration. Most materials have a recommendedshelf life of less than a year although this can generally be extended by

SOR3 12/19/98, 2:43 PM61


storage at temperatures around 5 °C. Guidance from the adhesive manufac-turer should be sought for storage on site under particular conditions.

3.3.4.2 Viscosity

The viscosity of adhesive materials can vary markedly with temperature,affecting workability, application quality and satisfactory wetting of theadherend surfaces. The property of thixotropy aids wetting of the substrateduring spreading, in adhesives which otherwise require a high viscosity. Forrelatively low ambient working temperatures it may be possible to obtainadhesives of a suitable viscosity from the manufacturer (e.g. summer andwinter grades).

3.3.4.3 Usable life

Where two component adhesives are mixed together in a can or container,there is a finite working life or pot life. This is related to the reactivity of theadhesive material itself, the mixed volume and the ambient temperature.The working life is extended at low temperatures and decreases at hightemperatures. It should be noted that the chemical reaction is exothermicand large volumes of mixed adhesive can become very hot. A typical usablelife might be in the range of 30–60min.

3.3.4.4 Wetting ability

The ability of an adhesive to wet out a substrate surface is fundamental toobtaining adhesion (see Section 3.4.2). Adhesives applied towards the endof their usable life tend to lack wetting ability.

3.3.4.5 Joint open time

This starts when the adhesive has been applied to the concrete, or FRP, orboth surfaces. It represents the time limit during which the FRP shouldbe applied to the concrete surface, or else the strength of the bond maybe reduced. This is a function of the degree of hardening and reactionwith the atmosphere itself. A typical open time might be of the order of30min.

3.3.4.6 Rate of strength development during cure

The cure time depends upon the type of adhesive and the ambient tempera-ture. The majority of cold curing epoxy formulations take 6–12h to achievea satisfactory degree of cure for handling purposes, with full cure beingattained in 24h or so (see also Section 3.5.6).

SOR3 12/19/98, 2:43 PM62

Materials 63

3.3.5 Mechanical characteristics

The properties of adhesive polymers are generally determined by theirinternal structure, but the blend of components in many commercial formu-lations prevents simple relationships from being drawn. A comparison ofthe tensile stress/strain behaviour of a range of epoxies is given in Fig. 3.4,whilst typical properties and characteristics are collected in Table 3.4. Thetable indicates the properties of Sikadur 31PBA, the adhesive used in theROBUST system of strengthening. A further discussion of the properties ofthis and other plate bonding adhesives is reserved for Chapter 11.

Adhesives possess a relatively low modulus which decreases with increas-ing temperature. For epoxies, Young’s modulus is typically between 1–10 GPa. Variations in temperature can transform materials which are toughand strong at 20 °C to ones which are soft and weak at 100 °C. The glasstransition temperature, Tg, denotes a marked change in the mechanicalproperties of a polymer. Above Tg the material will be rubbery, and belowit will be glass-like and stiff (Fig. 3.5). The Tgs of commercial buildingsealants are around �40 °C, whereas those of ambient temperature curingepoxies are around 40–50 °C. By warm- or heat-curing adhesives, their Tg

values are increased.Organic polymers all absorb moisture and one effect is to plasticise the

adhesive itself. This modifies its response to mechanical deformation ina manner analogous to a lowering of the Tg. Such effects are reversible,depending upon proximity to moisture in liquid or vapour form. Thus waterand heat have similar effects and this is illustrated in Fig. 3.6 and 3.7 for the

Table 3.4 Typical properties and characteristics of epoxy adhesives (2-part,cold cure)

Property (at 20 °C) Range Sikadur 31PBA

Shear strength (MPa) 15–35 28Tensile strength (MPa) 20–40 30Tensile modulus (GPa) 1–10 7Tensile strain to failure (%) 1–4 0.7Fracture energy (kJm�2) 0.2–1.0 0.4Glass transition temperature (°C) 40–60 58Coefficient of thermal expansion (10�6 °C�1) 30 30

Characteristics

Creep resistance ExcellentMoisture resistance ExcellentHeat resistance GoodCold or hot cure Cold/hotCure time Medium/longGap filling Yes

SOR3 12/19/98, 2:43 PM63

64S

trengthening of reinforced concrete structures

Figure 3.4 Typical tensile stress/strain curves for a range of epoxy adhesives.

SO

R3

12/19/98, 2:43 PM

64

Materials

65

Figure 3.5 The glass transition temperature.

SO

R3

12/19/98, 2:44 PM

65

66S

trengthening of reinforced concrete structuresFigure 3.6 Typical effects of temperature and moisture on the mechanical behaviour of an unmodified cold curingepoxy adhesive in shear.

SO

R3

12/19/98, 2:44 PM

66

Materials 67

Figure 3.7 Typical effects of temperature and moisture on themechanical behaviour of an unmodified cold curing epoxy in tension.

case of a stiff cold cure epoxy formulation with a relatively low strain-to-failure.

3.3.6 Testing

Testing represents a large and important subject in its own right and fallsoutside the scope of this book. A full discourse on the testing of adhesivesin bulk form and of testing adhesive bonded joints is given in Mays andHutchinson (1992). Some key test procedures are outlined elsewhere (Maysand Hutchinson, 1988).

The tensile stress/strain behaviour depicted in Fig. 3.4 was derived fromtests conducted using dumb-bell specimens. Values of tensile strength,modulus and strain-to-failure feature in Table 3.3. Reliable values of adhe-sive material shear strength can be derived from the thick adherend sheartest (TAST), as described by Mays and Hutchinson (1988, 1992), over arange of temperatures. Values for fracture energy can be obtained fromtests conducted using double cantilever specimens as described by Maysand Hutchinson (1992). One classical method of obtaining Tg is to usedynamic mechanical thermal analysis on a small beam of adhesive. A moreapproximate technique involves deformation of a larger prism of adhesivein order to determine the heat distortion temperature (HDT).

SOR3 12/19/98, 2:44 PM67


3.4 Adhesion and surface preparation

3.4.1 Introduction

A full general discussion on the topic of adhesion and surface preparationhas been given by many authors, including Mays and Hutchinson (1992).Particular discussion on adhesion and surface preparation of FRP surfacesis given by ASTM (1984), Clarke (1996) and by Hutchinson (1997). Somediscussion on adhesion and surface preparation of concrete surfaces isgiven by ASTM (1983), Sasse and Fiebrich (1983), Gaul (1984), Sasse(1986), Edwards (1987) and Hutchinson (1993).

It has been noted that the most important construction difficulty inbonded joint design is surface preparation (e.g. Clarke, 1996). A majorbarrier to the more widespread use of adhesive bonding in construction isa lack of understanding about adhesion, appropriate surface preparationtechniques and their effects on both initial bond strength and long termbond durability. Hot/wet environments are generally deleterious for adhe-sive bonds although there is little evidence to date that such environmentsare particularly harmful for FRP/concrete interfaces (see Chapter 6, Sec-tion 6.7). Nevertheless adequate surface preparation is still a vital prerequi-site for obtaining and maintaining joint integrity.

3.4.2 Adhesion and interfacial contact

Adhesives join materials primarily by attaching to their surfaces within alayer of molecular dimensions, that is, of the order of 0.1–0.5nm. The term‘adhesion’ is associated with intermolecular forces acting across an inter-face and involves a consideration of surface energies and interfacial ten-sions. Being liquid, adhesives flow over and into the irregularities of a solidsurface, coming into contact with it and as a result, interacting with itsmolecular forces. The adhesive then solidifies to form the joint. The basicrequirements for good adhesion are therefore intimate contact between theadhesive and substrates, and an absence of weak layers or contamination atthe interface.

Adhesive bonding involves a liquid ‘wetting’ a solid surface, which im-plies the formation of a thin film of liquid spreading uniformly withoutbreaking into droplets (Fig. 3.8). Fundamentally, the surface tension of theadhesive should be lower than the surface energy of the solids involved, inthis case, the treated surface of FRP and the exposed constituents of con-crete. Because of the similarity in adhesive and composite matrix com-position, values of surface tension and surface energy are similar. Bothcompositions contain polar molecular groups which are mutually attractiveand chemically compatible. Thus good adhesion is assured, provided that

SOR3 12/19/98, 2:44 PM68

Materials 69

Figure 3.8 Simple representation of non-wetting (a) and wetting (b). γL

is the surface free energy (surface tension) of a liquid, γS is the surfacefree energy of a solid and θ is the contact angle between solid andliquid.

contamination is removed by surface preparation. Some typical valuesof FRP surface energies and epoxy adhesive surface tension are given inTable 3.5.

The chemical and physical nature of the surface of concrete is complexand variable. The surface of this multiphase material may contain exposedaggregate, sand, unhydrated cement particles and cement gel, together withcracks and voids; the surface moisture content may also be variable. Surfacetreatments should remove significant contamination, cement-rich layersand traces of mould release agents. Mutual atomic or molecular attractionsbetween an adhesive and the exposed (and probably heavily hydrated)constituents may exist, but mechanical keying or interlocking into the ir-regularities and pores of the surface probably play a significant role too. Itis impossible to measure the surface energy of concrete, although it shouldbe possible to measure the energies of its constituents separately. Forexample, Table 3.5 indicates a value for silica.

Finally, the rheological characteristics of the adhesive (or a primer) arealso important and low viscosity materials are beneficial for penetratinginto, and binding together, surfaces such as concrete which are porousand which may be fractured at a microscale. The interaction of surfaceenergetics, tensions and rheology is a complex one.

SOR3 12/19/98, 2:44 PM69


Table 3.5 Typical values of surface free energies

Surface Surface free energy,γ (mJm�2)

Ferric oxide 13504

Aluminium oxide 6304

Silica 2904

Water 725

CFRP1 (peel ply) 64

CFRP2 (heavily abraded) 58GFRP3 (corona treated) 57GFRP3 (heavily abraded) 55CFRP1 (solvent degreased) 51Epoxy adhesive 455

GFRP3 (as moulded) 38

1 CFRP, vinylester matrix.2 CFRP, epoxy matrix.3 GFRP, polyester matrix.4 Under ideal conditions in vacuo; values measuredin air are 50–100mJm�2.5 Surface tension, rather than the surface freeenergy, of liquids is more usually quoted. Tensionand energy are numerically identical butdimensionally different.

3.4.3 Surface preparation

The purpose of surface preparation is to remove contamination and weaksurface layers, to change the substrate surface topography and/or introducenew surface chemical groups to promote bond formation. An appreciationof the effects of surface preparation may be gained from surface analyticalor mechanical test techniques. Surface preparation generally has a muchgreater influence on long term bond durability than it does on initial bondstrength, so that a high standard of surface preparation is essential forpromoting long term bond performance.

Surface preparation represents part of the bonding operation discussedin Section 3.5. However a detailed discussion is provided in this sectionbecause of the importance of understanding the effects of preparation onbond integrity.

3.4.4 Surface preparation of concrete

For concrete strengthening applications the material must be treated in situ,sometimes under less than ideal conditions. The plane of the surface(s) to

SOR3 12/19/98, 2:44 PM70

Materials 71

Figure 3.9 Schematic idealisation of concrete following surfacepreparation.

be treated (horizontal, vertical, overhead, etc.) has a large bearing on theselection of an appropriate method. The choice of method, or combinationsof methods, depends upon the costs, the scale and location of the operation,access to equipment and materials, and health and safety conditions.

In essence, the purpose of surface preparation is to remove the outer,weak and potentially contaminated skin together with poorly bound ma-terial, in order to expose small- to medium-sized pieces of aggregate (Fig.3.9). This must be achieved without causing microcracks or other damagein the layer behind which would lead to a plane of weakness and hence areduction in strength of the adhesive connection. Any large depressions,blow holes and cracks must be filled with suitable mortars prior to theapplication of a structural adhesive. This will ensure a relatively uniformbondline thickness in order to maximise the efficiency of shear stresstransfer. The basic sequence of steps in the process of surface preparationshould be:

• To remove any damaged or substandard concrete and reinstate withgood quality material.

SOR3 12/19/98, 2:44 PM71


• To remove laitance, preferably by grit blasting. Other techniques suchas wire brushing, grinding, bush hammering, acid etching, water jettingor flame blasting are not recommended for plate bonding applications.In particular, the use of pneumatic tools can cause significant damage tothe underlying concrete (Hutchinson, 1993).

• To remove dust and debris by brushing, air blast or vacuum.

Additional steps could include:

• Further cleaning, with a suitable solvent, to remove any remainingcontaminant.

• Drying the surface to be bonded, if necessary.• Application of an adhesive-compatible (epoxy) primer, if necessary.

The recommended method of laitance removal, by blasting, is fast, plantintensive and operator dependent. There exist a multiplicity of types ofblast media, media sizes, blast pressures and types of equipment. With drysystems, oil and water traps should be used to prevent contamination fromair compressors; dry systems may be open, or closed with vacuum recov-ery and recycling of the blast media. Generally a fairly microrough butmacrosmooth surface is generated, together with a lot of dust; particles ofblast media may also be left embedded in the surface. Clearly this dust anddebris must be removed prior to bonding. Wet blasting is another optionwhich overcomes some of the problems associated with the dust, but ofcourse a wet concrete surface is left which must be dried prior to bonding.Gaul (1984) suggests that the surface water content should be less than 4%prior to the application of coatings.

After preparation, the suitability of the surface should be checked usingthe pull-off test procedure described in Chapter 6, Section 6.7.2. The timelapse between preparation and bonding should be minimised in order toprevent any further contamination of the surface. Adhesion to damp andwet concrete surfaces was investigated in the ROBUST project, as de-scribed in Chapter 6, Section 6.7.2 and Table 6.3.

3.4.5 Surface preparation of fibre reinforced polymercomposites

For concrete strengthening applications the material may be treated off-site, enabling a variety of potential techniques to be used. However, rela-tively large areas of strip material will need to be treated in a reliable andconsistent way.

The surface of composite material may be contaminated with mouldrelease agents, lubricants and fingerprints as a result of the productionprocess. Further, the matrix resin may include waxes, flow agents and

SOR3 12/19/98, 2:44 PM72

Materials 73

‘internal’ mould release agents which can be left on the surface of a curedcomposite. Surface preparation not only serves to remove contaminationsuch as fluorocarbon release agents, but may also increase the surface areafor bonding, promote micromechanical interlocking and/or chemicallymodify a surface. Care must be taken to ensure that only the chemistry andmorphology of a thin surface layer is modified. It is important not to breakreinforcing fibres, nor to affect the bulk properties of the composite.

It is recommended in structural bonding that any random fibre mats andsurface veils are removed from the surface to ensure stress transfer directlyinto the main reinforcement fibres. Such a task is rather difficult to controlusing mechanical removal methods such as grinding or grit blasting. Forthis reason, careful consideration should be given to the fabrication of thecomposite material in the first place in order to make it ready for adhesivebonding with minimal disturbance. The main methods of surface prepara-tion for composites are solvent degreasing, mechanical abrasion and use ofthe peel-ply technique, and these methods are often used in combination. Afuller discussion of these and other techniques is considered by Hutchinson(1997). The effect of such treatments is to make the composite surface more‘wettable’, as shown in Table 3.5.

The basic sequence of steps in the process of surface preparation shouldbe either:

• To remove grease and dust with a suitable solvent such as acetone ormethyl ethyl ketone (MEK).

• To remove release agents and resin-rich surface layers by abrasion.This can be accomplished by careful grinding, sanding (e.g. ASTM,1984; EXTREN® Design Manual, 1989), light grit blasting using veryfine alumina (e.g. Bowditch and Stannard, 1980), or cryoblasting withsolid carbon dioxide pellets. The resultant surface should be textured,dull and lustreless.

• To remove dust and debris by solvent wiping.

or:

• To strip off a peel-ply layer. A peel-ply is a sacrificial layer about 0.2mmthick which is laid-up on the outermost surfaces of a composite materialand co-cured with it. During cure, the peel-ply layer becomes consoli-dated into the surface of the composite.

In the ROBUST system of structural strengthening, the CFRP stripswere pultruded with a clean scrubbed nylon peel-ply layer incorporatedinto their surfaces. This provides a uniquely practical surface preparationmethod such that unskilled operators can achieve a very clean and bond-able surface on site. This technique was first advanced for use for suchapplications by Hutchinson and Rahimi (1993).

SOR3 12/19/98, 2:44 PM73


There are several special considerations associated with peel-ply technol-ogy which are outside the scope of this book, but the subject is discussedmore fully elsewhere by Hutchinson (1997). Suffice to say that combina-tions of composite resin matrix, peel-ply materials and composite process-ing conditions need to be assessed carefully. When the dry peel-ply materialis pulled off, the top layer of resin on the FRP component is fractured andremoved, leaving behind a clean, rough, surface for bonding. The resultantsurface topography is essentially an imprint of the peel-ply fabric weavepattern (Fig. 3.10).

Finally, all FRP materials absorb a small amount of moisture due toambient humidity levels or proximity to a wet environment. In CFRP it isthe matrix resin which absorbs the moisture, but in GFRP and AFRP it isboth the fibres and the matrix resin which can absorb moisture. For ambienttemperature curing adhesive systems the presence of small amounts ofmoisture is unlikely to pose a problem although the EUROCOMP Hand-book (Clarke, 1996) recommends that the adherends be dried; this may not,of course, be practical! However with elevated temperature curing adhesivesystems the FRP components must be dried as far as possible (say to �0.5%moisture content by weight). This is because the heat curing process draws

Figure 3.10 Typical surface topography arising from woven peel-plymaterial.

SOR3 12/19/98, 2:44 PM74

Materials 75

moisture out of the adherend(s) and into the bondline, resulting in voidingand weakening of the adhesive itself.

The suitability of surfaces for bonding following any treatment can beassessed using wettability tests and mechanical tests. Single lap shear jointsof the type described in Chapter 6, Section 6.7.3 may be used to assessadhesion and joint strength.

3.5 The bonding operation

The satisfactory fabrication of joints relies upon the manufacture and pro-vision of high quality composite materials to specified tolerances, suitabledetailed designs and agreed methods of erection and completion of thework. Training of site operatives and their supervisors is essential. Moredetailed commentaries on the aspects detailed below are given by Mays andHutchinson (1992), Department of Transport (1994), Clarke (1996) andHutchinson (1997).

3.5.1 Handling, finishing and storage of materials

FRP components are relatively light and long lengths may be rolled up forconvenience. They should be stored with suitable protection and supportedoff the ground in well-ventilated dry conditions. A peel-ply surface layer, ifon both sides of the composite, will provide protection against scratching,chafing and damage to the bonding surface.

Any machining, trimming and drilling can give rise to problems ofsplitting, delamination and fibre pull-out. Thus FRP components must besupported rigidly with a suitable backing piece to enable these operations tobe carried out without causing damage. Large amounts of dust may begenerated which are best extracted directly by vacuum. Ideally, any sawn ordrilled surfaces should be sealed with a compatible resin to protect the fibrereinforcement from the direct effects of moisture in service.

Two-part epoxy adhesives are solvent free and simply require protectionfrom extreme conditions. Aspects of shelf life and the effect of temperatureon viscosity and application characteristics were discussed in Section 3.3.Cleaning solvents (and primers) contain flammable materials and they re-quire appropriate storage facilities at the place of fabrication.

3.5.2 Protection of the working environment

The working environment should be as dry and as clean as possible. Manyadhesives will not cure below about 5 °C, so that attention must be paid tocontrolling both the ambient temperature and the temperature of thecomponent surfaces. For example, a warm adhesive applied to a very cold

SOR3 12/19/98, 2:44 PM75


surface, even if properly prepared, may not adhere to it adequately becauseof the tendency to ‘skin’ due to thermal shock. Infrared heaters, spaceheaters and heating blankets can be used both to warm the componentsprior to bonding and to assist with curing postbonding. Dehumidifiers mayalso need to be used if the dew point is in excess of 10 °C.

The surfaces of components to be bonded must be dry and this can beachieved by proper attention to storage, an enclosed working environmentand by physically drying or preheating the components.


A high standard of surface preparation is essential, as discussed in Section3.4. However, surface preparation represents probably the most difficultpart of the bonding process to control and this often relates to the locationand scale of the strengthening operation. Thus simple, reproducible, surfacepreparation methods are required such as the use of the peel-ply techniquefor composite materials. It has been stated that the time elapsed betweencomponent surface preparation and bonding should be kept as short aspossible, principally to avoid subsequent contamination. Whatever themethod of surface preparation used, the resultant ‘qualities’ such aswettability, soundness and freedom from contamination can be measuredand compared against recommended criteria. A piece of adhesive tape canbe used to detect the presence of dust and debris. If moisture is picked upby absorbent paper pressed against the concrete it is likely to be too dampfor bonding.

3.5.4 Adhesive mixing, dispensing and application

Two-component adhesives must be mixed thoroughly and modern packag-ing generally removes the need for weighing or measurement, thus remov-ing the potential for errors. Resin and hardener components are frequentlyof dissimilar colours so that complete mixing can be judged when a uniformcolour has been achieved. With some systems supplied in tins, the hardenercomponent is poured into the container partially full of resin and mixing canthen take place. Alternatively, the resin and hardener may be supplied inco-axial or twin cartridges which can be mounted in hand-held guns; thepistons in the cartridges are advanced by a trigger mechanism and mixingtakes place within special disposable nozzles.

Where two-component adhesives are mixed together in a can or con-tainer, there is a finite usable life or pot life as described in Section 3.3.4.It was also noted that the viscosity, and therefore ease of application, ofadhesive material can vary significantly with temperature. Adhesives with

SOR3 12/19/98, 2:44 PM76

Materials 77

the correct characteristics at 20 °C may be impossibly viscous at 5–10°C, ortoo fluid at 30 °C.

It is important to spread the adhesive soon after mixing to dissipate theheat generated and extend its usable life. Adhesive may be applied either toboth concrete and the FRP, or to the FRP only. Common practice in platebonding is to spread the adhesive slightly more thickly along the centrelineof the FRP plates than at the sides of the plate. This can be achieved byusing a curved adhesive spreader tool to result in a convex adhesive layerprofile. This procedure reduces the risk of forming voids when pressing theplate against the concrete surface. Excess adhesive squeezed out of thebondline can be scraped away and a fillet formed at the plate edge.

3.5.5 Component fit-up and bondline thickness control

Good design and sequencing of the bonding operation should facilitate thepositioning of the plates, as well as controlling bondline thickness. Whereend-anchorages are used, precise positioning of the plates may be achievedby using anchor bolts placed through special plate end-tabs and intopredrilled holes in the concrete.

Thin FRP plates are so light that no temporary clamps or fixtures arerequired. The plates are simply pressed against the concrete surface usinghard rubber rollers and the ‘tack’ of the uncured adhesive is sufficient tohold the plates in position whilst it cures.

A target adhesive thickness of 1–2mm is recommended. A thick bondlinemay result in reduced joint strength and increased consumption of rela-tively expensive material. Conversely a very thin joint may result in ‘brittle’joint behaviour. Control can be achieved by the insertion of small thermo-plastic discs within the bondline or by the addition of glass spheres calledballotini to the mixed adhesive.

3.5.6 Curing

Curing and hardening can sometimes represent a significant process inscheduling and completing the bondline operation, giving rise to the needfor enclosure and artificial heating of the working environment (Section3.5.2). The actual rate of cure development depends upon the adhesiveformulation and the ambient temperature, with some scope often beingprovided by adhesive manufacturers who offer ‘normal’ and ‘rapid’ grades.A typical comparison is provided for Sikadur 31PBA, the adhesive used inthe ROBUST project, as shown in Fig. 3.11 and 3.12. A significant degreeof cure can be expected between 6h and 12 h over a range of temperaturesbetween 5–20 °C. Products will be almost fully cured after 24h.

SOR3 12/19/98, 2:44 PM77

78S


Figure 3.11 Shear strength (MPa) of Sikadur 31PBA normal and rapid grades from thick adherend shear test(TAST) joints cured and tested at 23°C.

SO

R3

12/19/98, 2:44 PM

78

Materials

79

Figure 3.12 Shear strength (MPa) of Sikadur 31PBA normal grade, from thick adherend shear test (TAST) jointscured and tested at different temperatures.

SO

R3

12/19/98, 2:44 PM

79


3.6 Durability and fire

Low temperatures generally have little effect on the strength and stiffnessperformance of composite materials other than to make the polymer lessflexible. This could result in a tendency to damage by fatigue and in extremecases it should be considered. High temperature on the other hand is moreproblematic, as all resins soften when heated sufficiently. It is thereforeimportant that a composite is designed and specified such that its designtemperature is compatible with its glass transition temperature (see alsoSection 3.3.5).

Carbon fibre is a conductor of electricity and composites in which it isused may be highly noble relative to some metals. If such materials arein contact, galvanic corrosion can take place, resulting in corrosion of themetallic element. Contact between the two materials must be avoided bythe incorporation of a suitable barrier.

Polymers absorb moisture to a certain extent and this has the effect ofaltering the properties of the polymer and of course a composite of which itis the matrix while wet. Mechanical properties and glass transition tempera-ture tend to be reduced as a function of the amount of water absorbed. Theeffect may be trivial in many circumstances but nevertheless should betaken into account by the designer.

The fire resistance of polymer composites ranges from ‘highly resistant’to ‘burns readily’. Fire resistance may be enhanced by high fibre content, ahigh level of filler in the resin mix and by the inclusion of fire retardantadditives if necessary. The fire performance characteristics which are mostappropriate for structural bonding applications are ‘surface spread of flame’and ‘ease of ignition’. Both must be designed to be appropriate for thecircumstances.

3.7 Painting

It is generally not necessary to paint composites, but if desired, paintingcan be carried out successfully; paint suppliers’ advice should be sought onspecific applications. Primers are available specifically for use with compo-sites which allow subsequent painting with cellulose and polymer lacquers.Two-pack polyurethane is a tough, long lasting paint system which does notgenerally need to be used in conjunction with a primer.

3.8 Summary

Structural adhesives have a long history of use in construction, with epoxyadhesives having been exploited for externally bonded steel plate reinforce-ment since the mid-1960s. Adhesive bonding represents the natural method

SOR3 12/19/98, 2:44 PM80

Materials 81

of joining together dissimilar materials such as concrete and polymer com-posites. Fibre reinforced polymer composites are being used increasinglybecause of their obvious advantages of high strength/weight ratio, designfreedom, ease of handling and corrosion resistance. The vast majority ofcomposites used in construction are reinforced with glass fibres, but themodulus and strength requirements for externally bonded reinforcementgenerally dictate the use of carbon fibre reinforced composites. Compositesproduced by the pultrusion and prepreg routes are of high fibre volumefraction, excellent mechanical properties in the fibre direction and of highreliability and quality. They must, however, be handled with care and dueprecautions should be taken when machining and drilling them.

Control of the adhesive bonding operation is crucial to the satisfactoryfabrication of all reliable and durable bonded joints. There are manyaspects which must be considered including the storage of materials, pro-tection of the working environment, surface preparation of the compo-nents, adhesive mixing, dispensing and application, joint fit-up, bondlinethickness control and curing. When bonding metals in particular, one of themost difficult aspects to control on site is surface preparation. By substitut-ing plastics for steel not only is the reliability of bonding improved, but byusing a peel-ply layer on the surface of the composite plates (as in theROBUST project) a reliable and practical site-based preparation techniqueis assured.

The durability of the materials and of the bonded interfaces is consideredin Chapter 6.

3.9 References

ASTM D2903 (1984) Standard Practice for Preparation of Surfaces of Plastics Priorto Adhesive Bonding, American Society for Testing and Materials, Philadelphia.

ASTM D4258 (1983) Standard Practice for Abrading Concrete, American Societyfor Testing and Materials, Philadelphia.

Bowditch M R and Stannard K J (1980) ‘Bonding GRP with acrylic adhesives’, inAdhesion 5, ed. K W Allen, London, Applied Science Publishers, Chapter 6.

Clarke J L (ed.) (1996) Structural Design of Polymer Composites, EUROCOMPDesign Code and Handbook, London, E & F N Spon.

Department of Transport (1994) Strengthening of Concrete Highway Structuresusing Externally Bonded Plates, BA 30/94, London, HMSO, Vol. 13, Section 3,Part 1.

Edwards S C (1987) ‘Surface coatings’, in Repair of Concrete Structures, ed. R T LAllen and S C Edwards, Blackie & Son, London, Chapter 10.

EXTREN® (1989) Design Manual, Morrison Molded Fiber Glass Company, BristolVA, USA, Section 13.

Gaul R W (1984) ‘Preparing concrete surfaces for coatings’, Concrete Int 6(7) 17–22.Hollaway L C (1993) Polymer Composites for Civil and Structural Engineering,

Glasgow, Blackie Academic and Professional.

SOR3 12/19/98, 2:44 PM81


Hutchinson A R (1993) ‘Review of methods for the surface treatment of concrete’,in Characterisation of Surface Condition, MTS Project 4, Report No 2 on thePerformance of Adhesive Joints, Dept. of Trade and Industry, London.

Hutchinson A R (1997) Joining of Fibre-Reinforced Polymer Composite Materials,Report 46, CIRIA, London.

Hutchinson A R and Rahimi H (1993) ‘Behaviour of reinforced concrete beamswith externally bonded fibre-reinforced plastics’, Proc. 5th Int. Conf. on StructuralFaults and Repair, ed. M C Forde, Engineering Technics Press, Edinburgh, Vol. 3,pp 221–228.

Mays G C and Hutchinson A R (1988) ‘Engineering property requirements forstructural adhesives’, Proc Inst Civil Eng 85(2) 485–501.

Mays G C and Hutchinson A R (1992) Adhesives in Civil Engineering, CambridgeUniversity Press.

Sasse H R (ed.) (1986) ‘Adhesion between polymers and concrete’, ProceedingsInternational Symposium in Aix-en-Provence, Chapman and Hall, London.

Sasse H R and Fiebrich M (1983) ‘Bonding of polymer materials to concrete’,RILEM Materials and Struct 16(94) 293–301.

SOR3 12/19/98, 2:45 PM82

Structural strengthening of concrete beams 135

135

5Structural strengthening of concrete beams

using prestressed plates

H N GARDEN AND G C MAYS

5.1 Introduction

The aim in prestressing concrete beams may be either to increase theserviceability capacity of the structural system of which the beams forma part or to extend its ultimate limit state. External prestressing can beapplied effectively by the use of unbonded tendons or bonded compositeplates. In both cases, it is essential that the existing structure can accommo-date the mounting of additional elements for the ends of the prestressingcomponents, but the use of tendons usually necessitates the installation ofadditional deviation supports for tendons to form the longitudinal profile ofthe tendons. In both cases, it will be essential to investigate fully the magni-tude and form of any local stresses which may be induced into the existingstructure as a result of the application of additional prestress.

The method of prestressing with externally bonded composite platescombines the benefits of excellent plate durability and structural improve-ment, due to the plate tension at the greatest possible distance from theneutral axis of the beam. It has been suggested that prestressing withcomposite plates is a more economical alternative to conventionalprestressing methods used in new construction (Triantafillou and Plevris,1991).

A significant advantage is gained by prestressing in segmental construc-tion (Trinh, 1990). This is due to the compressive strains generated byprestress at the joints between segmental units, these locations experiencinghigh tensile strain in the absence of prestress. This high tensile strain isassociated with high compressive strain at the top of a segmentally con-structed member, resulting in failure by concrete crushing. Depending onthe prestress magnitude, the likelihood of crushing at joints is either re-duced or eliminated. Leeming et al. (1996) reported the ability of externallybonded composite plates to restrict the opening of segmental joints.

The long term shortcomings of conventional prestressing using tendons isthat the tendon prestress does not remain at its initial value throughout thelife of the structure because various factors bring about prestress losses. The

SOR5 12/19/98, 3:32 PM135


principal losses are due to the following causes which contribute the pro-portions shown in brackets to the total prestress loss in post-tensioningconstruction:

• relaxation of the tendons (5%),• immediate elastic deformation of the concrete that occurs when pre-

stress is transferred into the beam (2–3%). Tendons that have alreadybeen prestressed will experience a loss of prestress due to the shorteningof the beam upon the prestressing of subsequent tendons,

• creep and shrinkage of the concrete under the compressive prestressover the service life of the structure (10–20%),

• slippage of the tendons at their end anchorages that occurs when theprestress is transferred into the anchorages,

• friction between the tendon and its duct in post-tensioned members, butthis being less of a problem in simply supported beams with fairly flattendon profiles (1–2%) but more severe in continuous beams with theirmore profiled tendons (10% or more).

Prestressing with externally bonded plates is also subjected to prestresslosses by immediate elastic deformation and long term creep of the con-crete, although friction losses do not apply. As with tendon losses due tosettlement and deformation at the end anchorages, bonded plates suffer aprestress loss due to the shear transferred through the adhesive and into theconcrete by the plate tension. This shear action is sufficient to fracture theconcrete even at low prestress levels so it is necessary to install anchoragesat the ends of the plate to resist this action.

One of the benefits of applying prestressed composite plates is that theeffect of the original internal tendon prestress can be restored. If the pre-stress loss is severe, the nominal tensile reinforcement will be unable tocontain concrete crack widths to within the serviceability limit and thepermissible applied load level will be reduced.

The application of externally bonded prestressed plates involves fourmain stages:

• the prestressing of the plate to that initial force which provides therequired long term prestress after losses have been taken into account,

• the bonding of the plate to the concrete member,• the installation of plate end anchorages that will resist the forces tending

to fracture the concrete when the prestress is subsequently transferredinto the member and when the member is subjected to an externallyapplied load,

• the transfer of the plate prestress into the concrete.

The prestressing force is not released until after the plate end anchorageshave been installed, in order to ensure that the cover concrete is not

SOR5 12/19/98, 3:32 PM136


damaged due to the high shear and peel stresses caused by the plate tension.Investigative work for the ROBUST project, undertaken at the Universityof Surrey, included interface tests on a concrete/CFRP (carbon fibrereinforced plastic) composite shear specimen in which the applied tensileforce was in the plane of the plate. Garden (1997) found that the maximumplate tension that could be transferred into the concrete, before surfacedelamination of the concrete occurred, was only 6% of the ultimate tensilestress of the plate material. In practice, much higher prestress levels, in theorder of at least 25%, will be necessary to achieve a significant improve-ment in structural stiffness and load carrying capacity of the concrete beam;it has even been suggested that a prestress of as much as 50% of the platestrength may be necessary (Meier et al., 1992)

5.2 Review of previous prestressing studies using

composite plates

Triantafillou and Plevris (1991) reported an analytical model developed todescribe the maximum achievable level of pretension which can be appliedto a composite plate so that the external strengthening system does not failnear the anchorage zones as the pretension is released. It was assumed thatfailure can occur either by horizontal cracking of the concrete above thefibre reinforced polymer (FRP) plate at the two end zones due to high shearstresses, or by yielding of the adhesive. The analytical model developed isused in a parametric study which suggests that the efficiency of the methodis improved by increasing the thickness of the adhesive layer, using athinner but wider plate of composite material, or by increasing the lengthover which the plate is bonded, all of these effects theoretically increasingthe level of stress which can be applied to the plate, prior to bonding, beforefailure occurs on release.

Section 2.4.3 of Chapter 2 has given a comprehensive review of theprevious prestressing studies using composite plates and it is clear from thisreview of the work reported in the literature that, whilst this extension ofthe FRP plate bonding technique is potentially beneficial, given the advan-tages inherent in prestressing in practice, it remains very much in its infancy.Most of the work carried out to date has merely demonstrated that failureof the system will occur on release of the prestress unless adequate anchor-age systems are provided at the plate ends. It has been demonstratedthrough the programme of beam testing, described in Chapter 4, that for theconfigurations tested and with shear span/beam depth ratios less than 4.0,failure of a non-prestressed system will occur by separation of the platefrom the beam when tested in flexure unless some form of anchorage isprovided. It therefore follows that anchorage is an even greater necessitywhen the plate is given an initial prestress. The provision of anchorages at

SOR5 12/19/98, 3:32 PM137


the plate ends reduces the shear deformation which occurs within theadhesive layer upon pretension release, thereby reducing the shear stressestransferred to the base of the concrete section and minimising the possibil-ity of premature failure. In the ROBUST investigations, therefore, plateend anchorages were used in all cases so that behaviour on application ofexternal load could be studied, as well as the initial response to pretensionrelease.

5.3 Prestressing technique employed in the

laboratory

The technique used to pretension the CFRP plates in the laboratory forboth the 1.0m and 2.3m long beams was as follows; the technique wasdifferent from that used on site (see Section 5.5).

Aluminium tabs were epoxy bonded at each end and on both faces of theCFRP plate. These tabs were used to provide stress distribution in the endregions and were 2mm thick, 130mm long for both beam sizes and full platewidth (80mm for the 1m long beams, 90mm for the 2.3m). Each end ofthe CFRP plate was then sandwiched between two steel plates 9mm thickwhich were predrilled and provided a jig into which the tabbed ends ofthe CFRP plate could be held with pins during drilling. After drilling, 126mm diameter bolts were located at each end and tightened to providefrictional as well as shear resistance to the prestressing force to be applied;the CFRP plate with the steel plates bolted on at each end wasthen loaded into a prestressing frame. The end reaction system is shown inFig. 5.1.

The sequence of procedures followed in preparing the prestressed beamsis shown in Fig. 5.2; methods of gritblasting and cleaning the concrete bondsurface, the utilisation of the peel-ply protection on the surface of thecomposite and the sprinkling of 2mm diameter ballotini to provide thecorrect thickness of adhesive were similar to those used in Chapter 4 forthe non-tensioned plates. After stressing the plate to the required level, theconcrete beam was lifted with jacks up to the level of the taut plate, Fig.5.2(a). A timber plank was then used to hold the beam at the correct heightwhilst the adhesive cured. Dead weights equivalent to a pressure of around7.5kNm�2 were applied to the upper face of the plate so that excess adhe-sive would be extruded, Fig. 5.2(b); in practice, a vacuum bag techniquemay be necessary to support the external plate during bonding. After theadhesive had cured, steel clamps were installed at each end of the beam toensure adequate anchorage of the plate ends upon release of the preten-sion. The load applied to the CFRP plate was then reduced to zero, thepretension being transferred to the beam by the cured adhesive bond layerand the plate was cut through to isolate the beam.

SOR5 12/19/98, 3:32 PM138


Figure 5.1 End reaction system for prestressing plates.

Figure 5.2 Sequence of procedures in preparing prestressed beams.

139

SOR5 12/19/98, 3:32 PM139


GFRP endplates 14mm thick and 40mm long, covering the full widthof the CFRP plate were used in the 2.3m and 4.5m beams. These werepredrilled and bonded in the correct positions, illustrated in Fig. 5.2(c).After this adhesive had cured and using the glass fibre reinforced plastic(GFRP) endplates as a jig, holes were drilled through the CFRP plateand adhesive into the concrete beam. Steel bolts were then bonded into theholes and allowed to cure, as shown in Fig. 5.2(d); it was necessary for thelength of bolt to extend well into the tensile zone of the reinforced concretebeam.

5.4 Results of laboratory tests for concrete beams

strengthened with prestressed plates in the

ROBUST programme

To illustrate the load/deflection and load/strain responses of reinforcedconcrete beams strengthened with pretensioned composite plates, typicalresults from the 2.3m long beams tested under the ROBUST programmewill be discussed. The four point load configuration adopted for thesebeams provided a 400mm long constant moment region in the centre of thebeam and each shear span was 450mm long.

The effects of strengthening the 2.3m beams with various plate tensionsare summarised in Tables 5.1 and 5.2, whilst the variations in overallmember stiffness are given in Table 5.3.

For all the 2.3m beams tested during the project, prestressing the com-posite plates prior to bonding increased the serviceability load but, as theload was governed by the concrete strength and not by the internal steel,the increase over the unplated case was small and pretensioning the plate

Table 5.1 2.3m Prestressing investigation; strengthening effects of plating;As/bd1 ratio was 1.18% and strength of concrete was nominally 50Nmm�2

(prestress as a percentage of the ultimate strength of plate)

Beam First Increase Service- Increase Yield Increasecracking over ability over load overload unplated load unplated (kN) unplated(kN) (%) (kN) (%) (%)

Unplated 13.0 — 47.2 — 76.5 —Plated 0% 15.0 15.4 53.3 13.0 100.0 30.7Plated 34.7% 22.5 73.1 53.3 13.0 115.0 50.3Plated 41.7% 34.0 162.0 53.3 13.0 124.0 62.1

1 As is the area of tensile steel rebars, b is the breadth of the beam and d is thedepth of the beam, respectively.

SOR5 12/19/98, 3:32 PM140


Table 5.2 2.3m Prestressing investigation; ultimate loads and ductilities; As/bdratio was 1.18% and strength of concrete was nominally 50Nmm�2

Beam Maximum load Increase over Ductility Proportion ofcarried (kN) unplated (%) unplated ductility (%)

Unplated 79.9 — 5.15 —Plated % 125.6 57.2 4.38 85.1Plated 34.7% 129.3 61.8 4.10 79.6Plated 41.7% 147.8 85.0 5.49 106.6

Table 5.3 2.3m Prestressing investigation; stiffening effects of plating. As/bdratio was 1.18% and strength of concrete was nominally 50Nmm�2

Beam Postcracking Increase over Postyielding Postcrackingstiffness unplated (%) stiffness stiffness(kNmm�1) (kNmm�1) retained (%)

Unplated 7.0 — — —Plated 0% 8.60 22.8 2.34 27.2Plated 34.7% 9.23 31.9 2.34 27.0Plated 41.7% 9.30 32.9 2.32 26.5

prior to bonding produced little benefit with respect to the serviceabilityload over the non-prestressed beam for the given concrete strength.

Figure 5.3 shows a typical relationship between the applied load andmaximum beam deflection under further tests at the University of Surreyfor the ROBUST programme. In this case the plate was initiallypretensioned to 25% of its ultimate strength, as indicated. The curvesreferred to as ‘unanchored’ and ‘bolted’ in this figure describe the absenceof plate end anchorage and the use of plate end anchorage bolts,respectively.

In all cases investigated, a considerable increase in the applied load tocause cracking in pretensioned plated beams was apparent and, from thecase considered in Fig. 5.3, a 100% increase over the non-prestressed beamwas achieved. The initiation and development of both flexural and shearcracking were much less in evidence for the pretensioned cases than for thenon-tensioned beams at comparable loads, demonstrating the ability of thepretensioned plate to limit cracking.

For the beams tested during the ROBUST investigations, it was observedthat the serviceability load was governed by the strength of the concrete(σu) and by the (As/bd) ratio. It should be noted that the results presentedin Tables 5.1, 5.2, and 5.3 corresponded to values of (σu) equal to 50Nmm�2

SOR5 12/19/98, 3:32 PM141


Figure 5.3 Deflection responses to applied load of beams with andwithout plate prestress (after Garden, 1997).

and values of (As/bd) equal to 1.18% whereas the equivalent values forthe results shown in Fig. 5.3 were 50Nmm�2 and 0.76%, respectively. Con-sequently, the increases in serviceability over the unplated cases were smallin the former case and relatively large in the latter.

From Fig. 5.3, the yield load was found to be 14kN for the unplatedbeam, 21kN for the plated unanchored non-prestressed case and 24kN forthe plated bolted non-prestressed cases, giving increases of 33% and 71.4%,respectively. Prestressing the plate was again found to increase the yieldload; an effective prestress of 25% produced a value of 31kN, an increase of121.4%.

The CFRP prestressing was also found to produce a moderate increase inthe maximum load carried by the plated beams compared with the non-prestressed ones. The nominally 50% prestressed plated case failed by platetensile fracture, whilst the nominally 25% prestressed plated beam failed byplate separation due to a vertical shear crack opening of the form describedin Chapter 4, Section 4.3. Therefore, the lower the ductility the more brittlethe failure, as the collapse mechanism changes from plate separation toplate failure in flexural tension.

A significant benefit of prestressing the plate is that the composite actionbetween the plate and the concrete, at the ends of the plate, improves as theprestress increases. This result is illustrated in Figs. 5.4 and 5.5, in which acomparison is made of the increments in plate tensile strain at variouspoints along the plate, under increasing applied load. Both figures representbeams with bolted plate end anchorage with a shear span/beam depth ratioof less than 3.4. However, it should be stressed that all pretensioned plates

SOR5 12/19/98, 3:33 PM142


Figure 5.4 Plate strain distributions with no initial plate prestress (afterGarden, 1997).

Figure 5.5 Plate strain distributions with a 40% initial plate prestress(after Garden, 1997).

require anchoring at their ends although non-tensioned plates above ashear span/beam depth ratio of 4.0 may not. The increased plate strain atthe plate ends indicates an improvement in the ability of the adhesive layerin the vicinity of the plate ends to transfer shear, although the precisemechanism by which this improvement comes about remains the subject offurther research.

SOR5 12/19/98, 3:33 PM143


Figure 5.6 Section strains without plate prestress (after Garden, 1997).

Figure 5.7 Section strains with 50% nominal plate prestress (afterGarden, 1997).

Figures 5.6 and 5.7 show, for two 4.5m long beams, the strain values atvarious heights in the plated beams for the non-tensioned and tensionedplates, respectively; the legends in the figures indicate the applied loads.The thick line, marked 0kN in Fig. 5.7, represents the strains in the sectionafter prestress transfer, but before external loading of the beam, for the casein which the plate was prestressed nominally to 50% of its ultimate strengthbefore bonding. Furthermore, when the applied load was 8kN, almost the

SOR5 12/19/98, 3:33 PM144


full depth of the beam was under a small compressive stress and on subse-quently applying external loads greater than 8kN, the neutral axis levelremained lower in the prestressed case than the non-prestressed one. Thehigh tensile strains, in the prestressed case, at the level of the compositeplate (Fig. 5.7) reflect the prestress tension applied to the CFRP compositeplate before the beam was externally loaded; the values at this level are thetotal strains in the composite plate.

A consistent feature of plated beams with end anchorage under flexuralloading is that the curve of applied load against plate strain after yield of theinternal rebars is common for non-prestressed and prestressed members(Garden et al., 1998); this is illustrated in Fig. 5.8. It is shown that thenon-prestressed member without plate end anchorage exhibited a lowerpost-yield stiffness than the non-prestressed beam with the bolted plateends. This is a typical result for beams loaded with a shear span/beam depthratio equal to or lower than 4.0 due to the stiffening influence of the plateend anchorages (Garden and Hollaway, 1997). This would indicate that thepost-yield stiffness of a plated beam with end anchorage is not governed bythe plate prestress and that the locus of yield points, with increasing pre-stress, defines a straight line. These features have been discussed by Garden(1997) in terms of the material properties and curvatures of plated sectionsunder the applied load.

During the investigations with non-prestressed plates, it was noticedthat sometimes the adhesive layer suffered cracking through its thicknessand along the interface with the composite plate, due to the propagation ofshear cracks through the depth of the concrete (Garden, 1997; Garden and

Figure 5.8 Total plate strains of beams with and without plateprestress (after Garden, 1997).

SOR5 12/19/98, 3:33 PM145


Hollaway, 1997). When the plate prestress is of the order of 20% of ultimate,this adhesive cracking can be considerably retarded or even eliminated. Inpractice, this characteristic would improve the durability of the system byreducing the moisture ingress at the level of the adhesive and its interfaces.

From a study of the number and distribution of shear cracks in the shearspans of beams loaded under symmetrical four point bending, Garden(1997) demonstrated that, for a given applied load with increasing plateprestress, the number of shear cracks reduces in relation to distance fromthe loading position. This observation reflects the improvement in durabil-ity that may be expected with a prestressed composite plate, compared witha non-prestressed one.

5.5 Results of field investigations of concrete beams

strengthened with prestressed plates in the

ROBUST programme

Two 18.0m long prestressed concrete beams were selected from the samesource as those used for the field investigations with non-prestressed plates,presented in Chapter 4. The lower three internal tendons in one of thesebeams were destressed, as described in Section 4.5, prior to the applicationof the prestressed plate. A four point bending load test, undertaken by theRoyal Military College of Science, using the loading configuration detailedpreviously in Section 4.5.3, was then undertaken within the elastic range toestablish the unplated member stiffness. In the second beam, the threelower tendons were destressed after the beam was plated; the prestressedplates were monitored to determine their increase in strain due to thedestressing of the internal tendons.

It was intended that both beams should be strengthened with threeprestressed CFRP plates in a single layer, each plate being 90mm wide by1.0mm thick bonded with Sikadur 31 PBA adhesive. The plates were 6.0mlong in the first beam and 15.8m in the second. Table 5.4 summarises thetest parameters.

Table 5.4 Test parameters for 18m beams (stressed plates)

Beam Overall plate Free/ Plate Maximum Maximum Monitoredno. dimensions Anchored length unplated plated during

(mm) plate ends (m) loading loading coring(kN) (kN)

7 270 � 1 A 6.0 62.5 1001

8 180 � 1 A 15.8 — 83 Yes

1 Beam failed under core holes by plate debonding.

SOR5 12/19/98, 3:33 PM146


In order to stress the plates it was necessary to bond a purpose-designedanchor block to each end of the plate. The purpose of anchorage blocks wasexplained in Chapter 4. The anchor blocks in the 18.0m beams were 530mmby 100mm by 15mm GFRP tabs, illustrated in Fig. 5.9. The anchor blockswere prebonded to the CFRP under factory conditions using the adhesive,9323-2B/A, manufactured by 3M. The plates were then transported to siteand prestressed to 30% of their ultimate tensile load using a prestressingdevice mounted on and reacting against the concrete beam, as shown in Fig.5.10. For safety reasons, the prestressing device and the bolted anchoragesystem were designed to resist a load equivalent to twice the ultimate tensilestrength of the layer of three plates.

Once stressed, the plates were pressed against fresh adhesive, spread to apredetermined thickness on the beam soffit, using a temporary timberpropping arrangement. This was necessary to accommodate the upwardcamber present in the beams which were already prestressed under theaction of the internal tendons; the propping system remained in place untilthe adhesive had cured.

This prestressing procedure was the first of its kind to be used successfullyunder site conditions. The work proceeded satisfactorily in all respects,except for the sudden and unexpected failure during stressing of one platein the second beam (no. 8). Subsequent examination of the failed platerevealed a weakening due to an overlap between successive peel-ply sheetswhich were entrapped within the body of the plate during manufacture.Important quality control lessons were learned from this work, of particularrelevance to prestressed plates which experience higher local stresses thannon-prestressed composites.

The curve of load against deflection for the plated beam, no. 7, is shownin Fig. 5.11, compared with the comparable unplated member. This showsthat there was a considerable increase in stiffness as a result of applying theprestressed plates, consistent with the smaller scale laboratory findings.This increase in stiffness correlated well with the upper bound stiffnesswhich was predicted using three-dimensional finite element analysis. Figure5.11 also shows the load–deflection behaviour of beam 5 which wasstrengthened with a similar configuration of non-prestressed plates (seeSection 4.4); the greater stiffness, arising from the plate prestress, is clearlyseen.

Beam 7 achieved an ultimate load of 100kN before failing by platedebonding under one of the core locations (see Fig. 5.12(a) and (b)). It isbelieved that this premature failure arose as a result of some damage to thetwo remaining prestressing tendons in the beam, which occurred during thedrilling operations to provide fixings for the plate prestressing device.

In beam 8, coring to destress the internal tendons was undertaken afterthe completion of the plate prestressing and plate strains were monitored asthe tendons were destressed. The result from the strain gauge immediately

SOR5 12/19/98, 3:33 PM147

148S


Figure 5.9 Plan of stressed anchor plate.

SO

R5

12/19/98, 3:33 PM

148


Figure 5.10 Stressing device mounted on beam.

beneath the core hole shows the development of tensile strain in the plateas the internal tendon was cut. As coring took place at other locations, thesepeak strains were redistributed along the length of the plate. There was nosign of any distress to the beam during this coring operation. The beam wasthen retested and gave the load–deflection response shown in Fig. 5.13. Thestiffness of this plated beam exceeded that predicted by the finite elementanalysis.

5.6 Observations

The prestressing studies reported previously in the literature, together withthe findings of the ROBUST investigative work, lead to a number of signifi-cant conclusions regarding the behaviour of beams strengthened with pre-stressed plates. The more important points may be summarised as follows:

• It has been shown that when bonding composite plates to concrete, theconcrete is unable to withstand peeling forces greater than about 5% ofthe plate strength and, therefore, it is necessary to provide an effectiveanchorage at the ends of the plate for prestressing forces greater thanthis value.

SOR5 12/19/98, 3:33 PM149

150S


Figure 5.11 Load deflection plots for beam 7.

SO

R5

12/19/98, 3:33 PM

150


(a)

(b)

Figure 5.12 Failure mode for beam 7.

SOR5 12/19/98, 3:34 PM151

152S


Figure 5.13 Load deflection plots for beam 8.

SO

R5

12/19/98, 3:34 PM

152


• It is essential that anchorage bolts are well bonded into the concrete toprevent concrete fracture after prestress transfer and during subsequentloading; the bolts must extend beyond the level of the internal tensilerebars.

• Composite action between the plate and concrete is maintained whenthe plate prestress is transferred to the concrete.

• Since the lower half of the beam is placed in compression by release ofthe prestress in the plate, flexural cracking will be much less extensiveand develop at a later stage under load compared to an identical non-prestressed beam. This effect of controlling cracks is significant inenhancing the durability of the reinforced concrete with respect tocorrosion.

• Prestressing will increase the serviceability load of a given member aslong as the stress carried by the internal steel is the governing factor;there is little benefit to be gained when the strength of the concretegoverns the serviceability load.

• Prestressing increases significantly the applied load at which the internalsteel begins to yield compared to a non-prestressed beam.

• Prestressing the plate prior to bonding also affects the mode of failure.The plate puts the lower half of the beam into compression throughoutits length, confining the concrete and reducing the extent of shear crack-ing which could initiate shear or shear step failure. As a result, thefailure will generally occur at the adhesive/CFRP interface or within thebottom layers of the concrete; this results in an increase in failure loadfor cases governed by a failure mode associated with shear.

• For a non-stressed plated beam where the shear span/beam depth ratiois below 4.0, the failure mode is likely to be due to plate separationinitiated at a wide shear crack. However, prestressing the plate preventsor reduces the opening of the shear cracks and a flexural failure is morelikely.

• Under a given applied load, the position of the neutral axis is lowerwhen the plate is pretensioned prior to bonding. Consequently, more ofthe concrete is in compression, resulting in a more efficient use of thematerial. Furthermore, the heights of the tensile cracks are reduced.

• With increasing applied load, a prestressed plate will delay the point atwhich the number and extent of shear cracks become significant.

• A prestressed plate with a smaller cross-sectional area can achieve thesame benefits of strengthening compared to that of a non-prestressedplate (i.e. improved cracking, yield and ultimate loads), thereby reduc-ing the material cost, although the installation costs will be greater.

• The field investigation demonstrated the feasibility of installing exter-nally bonded prestressed plates on site and has opened the way forfurther trials on bridge decks and other structures. Beams with pre-

SOR5 12/19/98, 3:34 PM153


stressed plates withstood higher ultimate loads and were stiffer thantheir counterparts with unstressed plates. The site tests also demon-strated that the provision of prestressed plates can act as a ‘safety net’against the failure of existing tendons in prestressed concrete elements;non-tensioned plates will provide the same effects, assuming that theyare adequately anchored at their ends. This failure may occur as a resultof corrosion, the reason why the bridge from which the full size testbeams were taken was demolished.

5.7 Concluding remarks

When strengthening a member, the level of prestress that can be appliedwill be limited by the tensile strength of the plate; tensile failure of the plateshould not precede either yielding of the internal steel or compressivefailure of the concrete to ensure adequate ductility. However, it should berecognised that, as the level of pretension is increased, so the stiffness of thestrengthened member is also increased, thereby reducing the tensile andcompressive strains. In this case, the member to be strengthened must haveadequate shear capacity for the enhanced ultimate load. This requirementis complicated by the fact that increasing plate pretension provides greatercrack containment, which increases the diagonal tension strength of thesection and may prevent the ‘peeling off’ failure mode associated with shear‘steps’; this latter may govern failure at lower prestress levels. From the testprogramme, it appears that the level of FRP pretension may also have to belimited by the strength of the plate end anchorages, by the horizontal shearstrength of the adhesive/FRP plate interface and by the bottom layers ofconcrete.

5.8 References

Garden H N (1997) The Strengthening of Reinforced Concrete Members UsingExternally Bonded Composite Materials, Doctoral Thesis, University of Surrey.

Garden H N and Hollaway L C (1997) ‘An experimental study of the strengtheningof reinforced concrete beams using prestressed carbon composite plates’, Proc. 7 th

Int. Conf. Structural Faults and Repair, Edinburgh, 1997, Vol 2, pp 191–199.Garden H N, Hollaway L C and Thorne A M (1998) ‘The strengthening and

deformation behaviour of reinforced concrete beams upgraded using prestressedcomposite plates’, J Mater Struct 31 (May) 247–258.

Leeming M B, Garden H N and Darby J J (1996) ‘The rehabilitation of deterioratedpost-tensioned beams by bonding with advanced composite plates’, Proc. FIPSymposium on Post-Tensioned Structures, London, 1996.

Meier U, Deuring M, Meier H and Schwegler G (1992) ‘Strengthening of structureswith CFRP laminates: research and applications in Switzerland’, in AdvancedComposite Materials in Bridges and Structures, 1st International Conference, edsK W Neale and P Labossière, Sherbrooke, Canada, pp 243–251.

SOR5 12/19/98, 3:34 PM154


Triantafillou T C and Plevris N (1991) ‘Post-strengthening of RC beams with epoxybonded fiber composite materials’, Proc. Specialty Conference on Advanced Com-posites Materials in Civil Engineering Structures, Nevada, 1991, pp 245–256.

Trinh J L (1990) ‘Structural strengthening by external prestressing’, Proc. NATOAdvanced Research Workshop on Bridge Evaluation, Repair and Rehabilitation,Baltimore, USA, April 30–May 2, 1990, pp 513–523.

SOR5 12/19/98, 3:34 PM155


156

6Environmental durability

A R H U T C H I N S O N A N D L C H O L L A W A Y

6.1 Introduction

It is envisaged that many applications of fibre reinforced polymer (FRP)strengthening would be outdoors and hence the durability of the concrete/FRP system under aggressive environments must be considered. The envi-ronmental resistance of any bonded assembly depends on the durability ofthe individual component materials, as well as on the bond between them.In the use of FRP materials for external strengthening of concrete, theindividual components are the concrete, the fibre reinforced polymer com-posite and the adhesive, which is usually an epoxy. In general, properlydesigned, compacted and cured concrete can be expected to show good longterm durability and should remain maintenance free for many years undernormal service conditions. The durability of concrete, either in a pre-stressed or reinforced form is possibly one of the most well studied subjectsin civil engineering because of its importance to everyday life and, conse-quently, it will not be discussed in this book. It should nevertheless be statedthat the cover concrete is likely to represent the weakest component inthe strengthened zone, so that the integrity of the strengthening system issomewhat dependent upon the properties of the concrete in shear and intension. In structural applications, the integrity of the adhesive bond andthe external FRP strengthening medium under adverse environmental con-ditions are the issues of prime importance and these will be discussed in thefollowing sections.

Progress in the field of plate bonding relies largely on demonstrating thelong term durability of the strengthening system under varying environ-mental conditions. As a bridge strengthening technique, the minimum re-quired life is 30 years. The environmental durability problems encounteredwhen steel plates are utilised as the strengthening medium are more intrac-table than those which may arise when composite materials are used. Thisis because of the difficulties of ensuring an adequate bond betweenthe adhesive and the steel, together with the possibility of electrochemicalcorrosion of the bonded surface.

SOR6 12/19/98, 3:35 PM156

Environmental durability 157

One of the most important requirements of an adhesive joint is the abilityto retain a significant proportion of its load-bearing capacity for longperiods under the wide variety of environmental conditions which are likelyto be encountered during its service life. The long term integrity of bondedjoints implies both chemical and mechanical durability under varying tem-perature, moisture and other environmental factors which, for externalpurposes, may include spray from de-icing salts or from the sea. Adhesivebonded joints with equivalent bond strength values in short term static testsmay differ markedly with respect to durability.

The measured residual joint strength after environmental exposure is afunction of change in the cohesive properties of the resin, the properties ofthe adherends and in the adhesion between the adhesive and adherend.Therefore, joint durability demands a three-fold consideration of the struc-tural integrity of the cured adhesive, the adherends and the environmentalstability of the interface. Joint design and material data should allow selec-tion of an adhesive type and FRP material which will themselves be suffi-ciently durable to withstand the service environment. The more complexproblem and the far more difficult to design against, is that of the environ-ment attacking the interfacial regions of a joint.

6.2 Environmental and service conditions

The durability of joints and particularly structural adhesive joints is gener-ally more important than their initial strength. Bonded joints used in a civilengineering environment may be subjected to a variety of service condi-tions. The normal service conditions to be considered are:

• temperature• moisture (humidity, liquid water, salt spray)• chemical attack (oil, fuel, chemical spills).

An abnormal hazard condition which may also need to be considered is thatof fire.

These service conditions should be considered in conjunction with theloading conditions which, for bridge strengthening, relate primarily to peakshort term static loading. Sustained loading (leading to creep), fatigue andimpact may also need to be considered.

It has been established that water, in liquid or vapour form, representsone of the most harmful environments for bonded joints (Kinloch, 1983).The problem is that water is found universally and the polar groups whichconfer adhesive properties make the adhesive inherently hydrophilic. Highenergy substrate surfaces (Section 3.4) are also generally hydrophilic.

Concrete itself is susceptible to the effects of moisture in a fairly predict-able way (Section 6.6). Of greater significance at this stage is that the

SOR6 12/19/98, 3:35 PM157


properties of the matrix resin in FRP materials, together with the propertiesof adhesives, are susceptible to the effects of heat and moisture. The resultof moisture absorption, which is reversible, is to lower the glass transitiontemperature (Tg) of these materials, leading to a change in their mechanicalproperties. The effect of elevated temperature is to reduce the strength andmodulus of polymers; the Tg of the adhesive is likely to be rather less thanthat of the matrix resin, so that the adhesive is the governing factor.

Adhesive bonded joints are generally affected by exposure to moistureand elevated temperature. In a well made joint where a sound bondhas been achieved, the main effect will be on the adhesive layer. A smallamount of moisture-induced plasticisation of the adhesive in highly stressedregions may actually be beneficial in reducing stress concentrations. How-ever, a small reduction in joint strength should normally be anticipated,in relation to the effects of environmental conditions on the adhesiveitself.

The resistance of connections to chemical attack depends upon thenature of the liquid and its effect on both the composite components andthe adhesive, if present. Alkalis can cause severe matrix resin softening witha consequent effect on any form of connection. Isophthalic polyesters pro-vide better resistance than orthophthalic polyesters in terms of alkalis andorganic solvents and are to be preferred for the majority of glass fibrereinforced plastic (GFRP) components. FRP components made with vinylester resins are better still, but a little more expensive. Epoxy materials,both as the matrix resin and as adhesives, are regarded as very inert in acidsand alkalis.

The resistance of joints to the effects of fire implies consideration ofthe entire FRP structure, and a useful commentary is contained in theEUROCOMP Design Code (Clarke, 1996). FRP materials and adhesivesare very poor conductors of heat, which is an advantage over metals, butthey can also possess a large coefficient of thermal expansion in directionsthat do not have a significant amount of continuous fibre reinforcement.The effect of dimensional changes of the components and joints directlyaffected by fire or heat should therefore be considered. Adhesives areweakened by the influence of elevated temperature and may char or burn ifexposed directly to fire.

6.3 Factors affecting joint durability

It is inadvisable to discuss joint durability without first reviewing the gen-eral behaviour and characteristics of bonded joints. A bonded joint repre-sents a layered system comprising different materials and interfaces, all ofwhich respond in different ways to an externally applied load or change inenvironmental conditions. It follows that a complete understanding of the

SOR6 12/19/98, 3:35 PM158


behaviour of a bonded joint under load is not a simple matter (Mays andHutchinson, 1992).

A problem with bonded joints is that much of the load is transmittedthrough edge zones, and it is these which potentially come under environ-mental attack first. In fact, the load is progressively borne by the innerregions of the joint although, nevertheless, it is often the most highlystressed regions which are under the greatest amount of environmentalattack. However, if the bonded area is sufficient to enable stress redistribu-tion within the adhesive layer, any changes in adhesive or cohesive proper-ties will not compromise the integrity of joints of a suitable geometricalconfiguration. The large areas involved in plate bonding imply that fewproblems may be anticipated, provided that a high degree of care has beenexercised in the design, specification and fabrication of such joints.

The main factors which influence joint performance and durability are:

• adherend type and nature• porosity/permeability of adherends• pretreatment• surface condition following pretreatment• primer type (if applicable)• moisture content of adherends at the time of bonding (concrete and

FRP)• adhesive type/cure cycle• postcuring of joints (if applicable)• joint configuration and exact geometrical details• exposure conditions• duration of exposure• imposed stress.

It will be appreciated that changes in any of the above factors can giverise to variations in joint behaviour.

One of the most important factors in joint durability is the environmentalstability of the adhesive–adherend interface. This is dictated by the type ofadhesive, the nature of the adherends and their surfaces. Changes inthe adhesive and the adherend can be allowed for; changes in adhesion areless easy to estimate. Thus, optimisation of surface conditions andpretreatments often represents the key to maximising joint durability. Thesurfaces of both concrete and FRP materials are relatively stable and, ifproperly prepared (see Section 3.4), durable bonds with epoxy adhesivesare formed relatively easily. This is in contrast to the situation with steelplate bonding, for which an adequate standard of surface preparation formild steel surfaces is quite hard to achieve in practice. Further, the surfaceof steel is fairly unstable, especially in the presence of water, such thatbonds to the oxide layer are susceptible to degradation. The substitution of

SOR6 12/19/98, 3:35 PM159


FRP materials for steel for strengthening applications is motivated in alarge part by the assurance of superior bond integrity.

The adhesive system selected is clearly very important (Section 3.3).Generally a two-part cold curing paste–epoxy material is used without aprimer. There are, however, occasions when a primer may be required forconcrete surfaces to ensure satisfactory adhesion. Conceptually, adhesivesrepresent natural candidates for joining FRP materials because they areoften similar in composition and nature to the composite matrix resin. Thefundamental concepts involved in adhesive selection are that it should:

• adhere well to the surfaces involved• exhibit low permeability to water• possess appropriate physical and mechanical properties.

The quality of the FRP material itself should be high for several reasons,as discussed in Section 3.2. These include the need for reproducible andpredictable properties (for design and prediction purposes), flatness (toensure uniform bonding and bondline thicknesses) and excellent consolida-tion (to reduce permeability and mechanical weakness). A poorly madecomposite will give rise to durability problems sooner or later.

Joint design, as discussed in Chapters 4, 5, and 8, has an importantbearing on joint durability. Among the important concepts are to allow fora large bond area and to avoid unacceptably large stress concentrations inthe joint. The synergistic effects of high temperature, excess moisture andapplied stress normal to the bondline are undoubtedly detrimental. Jointdesign should therefore seek to minimise the buildup of stress concentra-tions which give rise to indirect peel and cleavage loads at adherend/adhe-sive interfaces.

Finally the bonding operation, including protection of the working envi-ronment, is important (Section 3.5). Trained operatives working underskilled supervision should ensure that surface preparation, adhesive appli-cation, temporary clamping arrangements and adhesive curing details arehandled adequately. Poor control of the bonding operation generally mani-fests itself subsequently in joint performance and durability problems.

6.4 Environmental durability of adhesive bonded

joints

6.4.1 General observations

Experience with structural adhesive bonding has shown that the mechanicalproperties of bonded joints often deteriorate under warm and wet condi-tions (Bodnar, 1977; Kinloch, 1983). This is particularly so if the joints

SOR6 12/19/98, 3:35 PM160


comprise either high energy substrates such as metals, glasses and ceramicsor else permeable substrates such as concrete and timber (Venables, 1984;Mays and Hutchinson, 1992). Further, failure at the adhesive/substrateinterface, rather than failure within the adhesive layer itself, is commonlyfound only after environmental exposure.

With joints involving polymeric adherends, there are separate considera-tions relating to whether the adherends are thermoplastic or thermosettingin nature. Bonds to thermoset matrix composite surfaces, such as GFRPand carbon fibre reinforced plastic (CFRP), are stable in the presence ofmoisture; the effect of moisture is primarily to cause some plasticisation ofthe cured matrix resin. This may or may not affect the mechanical proper-ties of the material and therefore of a bonded joint, depending on the fibrelay-up and mode of loading (e.g. Section 6.6.2).

6.4.2 Diffusion and absorption of water

Water possesses special properties which can be related to its molecularstructure and which govern the way its molecules interact with each otherand with other substances. The polarity and ability of a water molecule toform hydrogen bonds makes it a universal solvent, allowing it to dissolve,soften or swell organic substances whose molecules contain sufficient polargroups, such as epoxide. Thus, polar adhesives are naturally hydrophilicwhereas non-polar plastics, such as polyvinyl chloride (PVC) and poly-thene, are not. The solubility of water in epoxies is of the order of a fewmass percent, and the coefficient of diffusion of water at 20 °C is around10�13 m2 s�1 (Mays and Hutchinson, 1992). The permeability of a material isgiven by the product of diffusion and solubility (Comyn, 1983).

For moisture to affect an adhesive joint it must first enter the joint eitherby ‘wicking’ along the interface between the adhesive and adherend, bydiffusion into the adhesive and adherends, or through cracks and crazes inthe materials involved. Wicking may be significant if appropriate surfacetreatments have not been carried out so that initial adhesion is minimal.However diffusion will be a dominant mode of entry in concrete andpolymer composite adherends, together with the presence of cracks inthe concrete.

Diffusion, whilst controlled primarily by concentration differences, isinfluenced significantly by temperature; the higher the temperature, thefaster the rate of diffusion. Heat-cured polymers generally possess fairlyrigid molecules which reduce their level of molecular motion and hinderwater diffusion. The implication is that a pultruded composite matrix mate-rial, with a Tg � 100 °C, should be less permeable than a cold-cured epoxymaterial with a Tg in the range 50–60 °C.

SOR6 12/19/98, 3:35 PM161


6.4.3 Processes involved in joint degradation

In considering the behaviour of a bonded joint, it is useful to separate theeffects of moisture, temperature and stress on the:

• adhesive material• adherend material• adhesion between the adhesive and adherend materials.

For the case of using FRP materials to strengthen existing concretestructures it is necessary to consider the effects of these agents separately onthe adhesive, the concrete, the polymer composite, the adhesive/concreteinterface and the adhesive/polymer composite interface. In general, themain processes involved in the hydrolytic deterioration of bonded joints are(Comyn 1983, 1985):

• absorption of water by the adhesive• absorption of water by the adherends• absorption of water at the bonded interface(s) through displacement of

adhesive• corrosion or deterioration of the substrate surfaces(s).

For the case of FRP materials bonded to concrete it is only necessary toconsider in detail the first two points; absorption of water at the bondedinterfaces, and corrosion or deterioration of the substrate surfaces can bedisregarded. If FRP is bonded to other substrates such as timber, cast ironor steel then the considerations are rather more complex (e.g. Mays andHutchinson, 1992).

The general effects of moisture on polymeric materials are outlined inSection 6.2. These may be summarised by stating that moisture will causesome plasticisation of both the adhesive and the polymer composite matrixmaterial, and that such effects are reversible. The magnitude of such effectsis dependent upon the particular formulations of polymer involved and theconditions under which they have been cured.

In GFRP it is known that moisture can attack the surface of glass fibresleading to corrosion, and that adhesion between the fibres and matrix resinmay be reduced (see Section 6.6.2). Clearly this will lead to a reduction inthe mechanical properties of the material and therefore also of the bondedjoints that are made with it. However generalisations are dangerousbecause of the influence of the fibre sizing, the orientation of fibres withrespect to applied loads and the quality and fabrication parameters associ-ated with its production. In aramid fibre composites (AFRP), moisture maybe absorbed by the fibres themselves, leading to a loss of fibre propertiesand therefore composite material properties; in turn this may lead to

SOR6 12/19/98, 3:35 PM162


changes in the behaviour of bonded joints made with AFRP. No suchdifficulties have been reported for CFRP to date.

6.5 Procedures for assessing environmental effects

on materials and on bonded joints

6.5.1 General remarks

A number of standard test procedures exist for assessing the effects ofenvironmental conditions on materials such as concrete, polymer compos-ite materials and on adhesives (in bulk form). Whilst a similar number ofstandard test procedures exist for assessing the behaviour of adhesivebonded joints, very few of these can provide useful information on environ-mental effects. Hardly any procedures actually provide quantitative datafor the reasons outlined in Section 6.5.3.

The selection of laboratory exposure conditions presents a significantdilemma. Accelerated testing is commonly achieved by increasing the tem-perature, proximity to moisture and the imposition of load. However, onlymoderate increases in temperature above likely operating maxima shouldbe used in order to prevent degradation mechanisms taking place whichwould not occur in practice. For example, the mechanical properties of theadhesive material will be reduced above its Tg, water uptake will increasemarkedly above Tg and hydrolysis of some adhesive materials may occur.Freeze–thaw cycling is favoured for construction applications in particular,where thermal shock or freezing of clustered water molecules may give riseto joint failures either directly or indirectly. BS EN 29142 (1993) describesseveral single and multivariable atmospheric ageing regimes.

When considering the use of accelerated testing involving elevated tem-peratures, it is advisable to postcure joints constructed with cold cure adhe-sives. This is to prevent the exposure environment itself from postcuring theadhesive and altering its mechanical properties. Clearly it is necessaryto maintain a known base line of joint performance and this can only beachieved by being able to eliminate the effects of temperature itself onbehaviour. The joints described in Section 6.7.3 were postcured at 50 °Cprior to exposure.

It is most useful to collect information on joints exposed to naturalweathering conditions because some workers have found that natural expo-sure is worse than laboratory ageing. This may be related to the effectsof cyclic environmental effects. Exposure in hot/wet climates gives rise tofaster joint degradation than exposure in more temperate climates. Seacoast exposure is also generally more demanding than exposure to indus-trial environments.

SOR6 12/19/98, 3:35 PM163


6.5.2 Experimental considerations – adhesive andcomposite material behaviour

Bulk material characterisation represents a useful adjunct to durabilitytrials on joints, enabling environmental effects on the adhesive and compos-ite materials themselves to be separated from those on adhesion and onoverall joint behaviour. It has already been stated that the influence ofwater (and heat) on the adhesive and on a composite matrix resin is gener-ally reversible; water uptake is accommodated largely by swelling and itseffect is to reduce Tg. The modulus and strength of adhesives and matrixresins are lowered by plasticisation although fracture toughness and ductil-ity generally increase (Figs. 3.6 and 3.7).

Tests may be carried out using adhesive materials in bulk form whichhave been cast to shape. Typical geometrical configurations include tensiledumb-bells, blocks for compression, rods or strips for torsion pendulumtests and films for water sorption experiments. Quantitative information onshear stress–strain behaviour can also be obtained from elaborate jointswhich ensure ‘pure’ shear deformation in the bondline. Fracture mechanicsspecimens employing cantilever beam arrangements can be used to obtainvalues of tensile opening (mode 1) fracture energy, G1, in order to maptoughness as a function of environment (see Section 3.3.6).

Short term tests on fibre reinforced polymer composite materials are welldocumented in various Standards, and many of these can be extended intoenvironmental exposure trials. For example, flat rectangular test pieces canbe subjected to tensile and bending stresses in order to determine strengthand modulus following periods of accelerated ageing. Small pieces can alsobe used for water sorption experiments.

6.5.3 Experimental considerations – bonded joints

In the majority of test joint configurations, the adhesive bondline stressesare far from uniform. Failure loads are therefore related to stress concen-trations of the ends of a joint, so that if the bonded area is sufficient toenable stress redistribution within the adhesive layer due to changes in theadhesive material properties, apparent increases in joint ‘strength’ and‘toughness’ may occur initially.

A number of competing mechanisms are taking place and the effects ofthese are more noticeable in smallish joints unless initial adhesion is verypoor. Test joints using small bonded areas are preferred for durabilitytesting in order to minimise experimental timescales. Thus the small jointswhich are generally used do provide fairly rapid information, but the quitedramatic changes in joint behaviour which sometimes occur should notnecessarily cause undue alarm. Small joints can provide useful comparative

SOR6 12/19/98, 3:35 PM164


information on, say, the behaviour of different bonding systems or theeffectiveness of different surface treatments. However the size effects in-herent in testing mean that no direct correlation with the behaviour of largereal-scale joints can be made.

Appropriate comparative tests of adhesion and bond durability shouldsubject the interface to tensile, peel or cleavage forces. Thus pull-off tests,peel tests, fracture energy tests and, ironically, lap-shear tests are routinelyemployed. The choice depends largely upon the nature of the adhesives andsubstrate materials involved.

Special consideration needs to be given to the modes of loading imposedon joints involving both concrete and polymer composite materials. Con-crete represents a brittle substrate material of low tensile strength, such thatdirect loading of joints made with it tends to result in premature failure ofthe concrete. Thus joint configurations which put the concrete into com-pression are often recommended to avoid substrate failure. Such configura-tions include slant shear (Fig. 6.1) and compressive shear tests (Fig. 6.2),resulting in a measure of the resistance of a bondline to a combinationof shear and compression. Three- and four-point bending tests are alsoemployed (Fig. 6.2), resulting in a measure of average shear or shear andtension resistance. Notwithstanding the above, the partially cored pull-offtest (Fig. 6.3a) is used routinely for testing the bond integrity of concreterepair materials. It has also been used in modified form to assess bondsbetween epoxy adhesives and concrete surfaces (Fig. 6.3b). Results usingthis procedure are outlined in Section 6.7.2.

Polymer composites are also relatively brittle materials. The distributionand orientation of the fibre reinforcement has an enormous influence on the

Figure 6.1 Different types of slant shear test configurations for jointsinvolving concrete substrates.

SOR6 12/19/98, 3:35 PM165


Figure 6.2 Different types of bending and compressive shear tests forjoints involving concrete substrates.

load-carrying capacity of joints. In general, the through-thickness strengthof composite materials is low because of the layered compositionand, sometimes, resin-rich layers. Bonded joints are therefore prone tointerlaminar failure. Laboratory experience has shown that single lap shearjoints and wedge cleavage joints are satisfactory for use where a highproportion of fibres are parallel to the principal direction of loading. Singlelap shear data is given in Section 6.7.3.

6.6 Effect of environment on the component

materials used in the ROBUST system

To verify that the ROBUST material systems could maintain stability,separately and in combination under hot/wet environments, the consortiumexperimentally studied exposed samples of Sikadur 31PBA adhesive andthe carbon fibre/vinylester polymer composite strengthening plates. Rein-forced concrete (RC) beams 0.8m long strengthened with these materialswere also manufactured and their specifications and results are discussed inSection 6.7.4.

SOR6 12/19/98, 3:35 PM166


Figure 6.3 Partially cored pull-off tests.

SOR6 12/19/98, 3:36 PM167


The test and exposure conditions adopted for the Sikadur 31 PBA adhe-sive, composite materials and plated beams were as follows:

• Thermal cycling of Sikadur 31PBA adhesive, GFRP and CFRP compo-site plate coupons and unloaded CFRP plated RC beams, betweentemperatures of �20 °C and �50 °C, represented warming and coolingeffects. One cycle took 24 h with the rising temperature taking 5h andcooling temperature taking 17 h; at the extreme temperatures there wasa dwell time of one hour. During exposure, the natural moisture inthe air condensed on to the material coupons of adhesive, CFRP andGFRP. After each of the exposure periods of 0, 50 and 180 thermalcycles the specimens were loaded to failure to assess the influence of thefreeze–thaw environment.

• Sikadur 31PBA adhesive, GFRP and CFRP composite plate couponswere exposed to a warm, moist atmosphere of 30 °C and 100% relativehumidity. They were tested after 0, 50 and 180 days; in each case thematerial was tested 24h after its removal from the humid atmosphere.

The results from these tests are described in Sections 6.6.1 and 6.6.2, andsummarised in Tables 6.1 and 6.2.

6.6.1 Sikadur 31PBA adhesive

The general properties and characteristics of this two-part cold curing filledepoxy adhesive are discussed in Section 3.3, together with the effects ofmoisture on its properties. In general, the type of structural adhesive usedcan affect both the rate and degree of environmental attack on bondedjoints (Minford, 1983). A bonded joint may be appreciably weakened ifthe adhesive is chemically attacked to any significant extent by the serviceenvironment. The detailed chemistry of the adhesive also appears to influ-ence the stability of the interfacial regions due to the formation of morestable intrinsic interfacial forces. Furthermore, the physical impact whichthe service environment has on the adhesive is dependent on the composi-tion of the adhesive.

The more highly filled the adhesive, the lower should be its long termwater absorption (Tu and Kruger, 1996). The absorption of moisture tendsto accelerate time-dependent processes by lowering the Tg, thereby reduc-ing the performance of the adhesive at high temperatures. As such, mois-ture combined with heat has a particularly unfavourable influence on theadhesive. However, whether plasticisation of the adhesive layer by theingress of the water actually affects the strength of the bonded joint is hardto predict since, for many joint configurations, a decrease in the adhesivemodulus may decrease the stress concentrations in the joint and lead to anincrease in joint strength. Similarly, the toughness of joints subjected to

SOR6 12/19/98, 3:36 PM168


fracture often increases somewhat because of greater plastic deformationand enhanced crack-tip blunting properties within a plasticised matrix(Ripling et al., 1971; Kinloch and Shaw, 1981; Kinloch, 1982). Cohesivestrength may, however, eventually be reduced sufficiently to offset theincreased toughness (Hutchinson, 1986).

The mode of failure of bonded joints when initially prepared is usually bycohesive fracture in the adhesive layer, or possibly in the substrate materialsif these are particularly weak. However, a classic symptom of environmen-tal attack is that after such exposure, the joints exhibit some degree ofinterfacial failure between the adhesive and the adherend (Brewis et al.,1982). The extent of such failure increases with time of exposure to thehostile environment. Whether the failure path is truly at the interface, orwhether it is within a boundary layer of the adherend or adhesive remainsa matter of debate. Several authors (e.g. Brockmann, 1983; Kinloch, 1987)have emphasised that the structure of the cured adhesive adjacent to theadherend surface differs from that of the bulk, because of the influence ofsurface morphology and chemistry on the initial wetting and absorption ofadhesive. The inference is that this weak boundary layer of adhesive may beless densely cross-linked and/or have a lower concentration of filler parti-cles than that of the bulk material, and may therefore be more susceptibleto hydrolytic destruction. The rate of interfacial transport of water couldalso be somewhat higher than that through the bulk material.

To enable some material characterisations of the adhesive used in theROBUST system to be obtained, bulk coupons were exposed to the hot andwet environmental conditions described earlier.

Table 6.1 shows the adhesive characteristics after the three periods of thetwo types of exposure; the modulus is given as the initial tangent modulus

Table 6.1 Adhesive material characteristics after environmental exposure(Garden and Hollaway, 1997)

Exposure type and Modulus of Tensile strength Ultimate strainduration elasticity (GPa) (MPa) (microstrain)

Thermal cyclingUnexposed (control) 9.0 (� 21.1%) 26.3 (� 27.0%) 4300 (� 46.5%)50 thermal cycles 11.3 (� 0.5%) 29.6 (� 10.8%) 3467 (� 18.3%)100 thermal cycles 11.5 (� 1.1%) 31.4 (� 9.7%) 4302 (� 17.6%)180 thermal cycles 11.8 (� 19.7%) 38.2 (�13.3%) 4333 (� 19.2%)

Humidity exposureUnexposed (control) 9.0 (� 21.1%) 26.3 (� 27.0%) 4300 (� 46.5%)50 days 7.8 (� 1.2%) 13.4 (� 5.3%) 3663 (� 22.5%)100 days 6.4 (� 10.2%) 16.2 (� 7.2%) 5284 (� 21.5%)180 days 5.7 (� 19.7%) 15.7 (� 2.5%) 7980 (� 20.8%)

SOR6 12/19/98, 3:36 PM169


value at zero stress. In the thermal cycling environment the stiffness of thematerial increased in value, indicating a continuing polymerisation process.However, the negative influence of the humidity atmosphere on the adhe-sive indicates an increase in ductility associated with the plasticisation ofthe material. The ultimate strain increase indicates moisture uptake withplasticisation occurring and the experiment does reflect the serious effect ofa high humidity environment. However it should be noted that plate bond-ing systems, whether for strengthening of bridge or building structures, willnot generally be exposed to such extreme environments.

6.6.2 FRP adherends

Joints involving FRP adherends are far less susceptible to environmentalattack by water than are joints made with adherends with higher energysurfaces such as metals (see also Section 3.4). However, in most en-vironments, polymer composites show a degree of change with time. InSection 6.4.3 it was stated that the most important factors in inducing suchchange are moisture and natural weathering, and to these might be addedelevated temperatures. In addition, the effects of sunlight, particularlythe UV component, can have some influence on degradation. Separately,and in combination, these factors may all contribute to a deterioration inproperties.

The effects of the service environment on the mechanical, physical andchemical properties of polymer composites depend to a large extent on theefficiency with which the composite is prepared, in particular the quality ofexposed surface. However, the contribution of the polymer resin systemused is the most important factor in relation to durability. The stabilityof the polymer depends on the chemistry and conformation of the mol-ecules, and is strongly dependent on the cross-linked structure in thermo-setting polymers. Since the resins used to manufacture composite materialsand adhesives subsequently used to bond them are both polymeric, consid-eration of the effects of the environment are equally applicable to bothcases.

In civil engineering structures it should be mentioned that the compositewill not normally be heated to a temperature near to the limit of its me-chanical performance. This temperature is dependent upon the manufac-turing technique, but would typically be in excess of 140° for polyesters,vinylesters and epoxies. However, temperature exposure has three effectson polymer composites. In conditions of constantly fluctuating tempera-tures, differences in the thermal expansion coefficients of the resin andreinforcing fibres could contribute to progressive debonding and weakeningof the materials. However, for well prepared composites (this invariablyimplies mechanical production) this is not generally a problem.

SOR6 12/19/98, 3:36 PM170


Unreinforced polymers have very high coefficients of thermal expansion,but the values are considerably reduced by the addition of fibres and fillers.The effect of temperature on the matrix properties is usually reversibleunless the Tg is approached. The resistance of materials to strain under loadis highly dependent on temperature because of the viscoelastic propertiesof the matrix. It must be stated that the above conditions are the worstscenario for composites and, although plate bonding which is undertakeninside buildings is not generally exposed to continually fluctuating or largevariations of temperature, these effects might be relevant to bridgestrengthening where externally bonded plates are exposed to diurnal andseasonal fluctuations.

As discussed for adhesives in Sections 6.2 and 6.6.1, moisture is usually amore significant factor than heat in causing deterioration of composites,since it is likely to have both chemical and physical effects on all of thecomponents individually and on their interaction. The effects of moisturedepend upon the specific polymer and fibre materials utilised in the manu-facture and design of the composite, as well as the duration and globalenvironment to which it is exposed. Moisture will be absorbed by thecomposite if the resin is sufficiently hydrophilic (Loos et al., 1981). Inaddition, the ingress of water through capillary channels, voids and atexposed cut surfaces will affect the rate and extent of degradation sincemoisture molecules destroy some of the chemical bonds in the resin. How-ever, this problem is mainly relevant to GFRP composites. The ROBUSTsystem uses CFRP in which the matrix is vinylester; this system has excel-lent resistance to corrosive environments and to moisture uptake (seeSection 3.2).

Moisture reduces the Tg of polymeric resins and has a plasticising effectwhich, in turn, affects the properties of FRP composites. However, changesin matrix moduli have little effect on longitudinal tensile strength andmodulus of the composite since the matrix plays only a minor role in theseproperties. In addition, the decrease in elastic modulus of the resin due toany uptake of water is reversible in that the modulus returns to its originalvalue when the moisture diffuses out. As the coefficient of thermal expan-sion of unidirectional carbon fibre/polymer composite of about 60–65% byweight of fibre is almost zero (carbon fibre has a negative coefficient ofthermal expansion), the ultimate tensile strength of this type of compositeis relatively insensitive to temperature in the range �73 °C to �107 °C,regardless of the moisture content of the material. The temperature rangeover which plate bonding would be utilised is well within these limits.Temperatures in the range �73 °C to �177 °C were observed by Shen andSpringer (1981) to have a negligible effect on the modulus of elasticity,regardless of the moisture content of the material (their moisture contentvaried from dry to fully saturated). It might also be supposed that moisture

SOR6 12/19/98, 3:36 PM171


will have little effect on the tensile properties if the fibres themselves areunaffected, as should be the case with carbon.

As stated in Section 6.2, composites composed of glass fibres in polyestermatrix can be vulnerable to environmental attack but their resistance doesincrease when going from orthophthalics to isophthalics to vinylester poly-mers (Norwood, 1994). Furthermore, a reduction in strength of the fibresunder atmospheric conditions has been observed within short periods oftime after they have been subjected to load. This static fatigue or stresscorrosion results from a chemical reaction between water vapour and thesurface of the glass that permits a pre-existing flaw to grow to criticaldimensions and bring about spontaneous crack propagation. The rate ofthis reaction is dependent on the magnitude and local stress conditions, aswell as the temperature, pressure and composition of the surrounding at-mosphere. ROBUST confirmed that CFRP composites were far superior toGFRP under these atmospheric conditions and consequently this systemconcentrated upon CFRP composite.

Table 6.2 shows the results of the CFRP and GFRP tensile tests followingexposure of the specimens to the thermal cycling and humidity regimes

Table 6.2 Composite material characteristics after environmental exposure(Garden and Hollaway, 1997)

Material type and Modulus of Tensile Ultimate strainexposure time elasticity (GPa) strength (MPa) (microstrain)

CFRP thermal cyclingUnexposed (control) 110.88 (� 4.2%) 1414 (� 7.6%) 12340 (� 6.0%)50 thermal cycles 113.40 (� 5.1%) 1386 (� 6.2%) 11827 (� 5.2%)100 thermal cycles 116.28 (� 3.4%) 1627 (� 7.9%) 14060 (� 7.1%)180 thermal cycles 121.44 (� 2.2%) 1734 (� 8.1%) 14348 (� 6.0%)

GFRP thermal cyclingUnexposed (control) 36.09 (� 2.9%) 955 (� 6.5%) 26185 (� 7.3%)50 thermal cycles 36.65 (� 3.5%) 1056 (� 5.2%) 28512 (� 5.8%)100 thermal cycles 37.62 (� 2.5%) 973 (� 5.8%) 28104 (� 8.8%)180 thermal cycles 37.17 (� 4.8%) 1004 (� 4.8%) 19350 (� 7.3%)

CFRP humidity exposureUnexposed (control) 110.8 (� 4.2%) 1414 (� 7.6%) 12340 (� 6.0%)50 days 109.95 (� 3.8%) 1587 (� 8.5%) 13967 (� 5.2%)100 days 111.77 (� 4.5%) 1476 (� 3.7%) 13471 (� 3.7%)180 days 114.57 (� 3.5%) 1609 (� 8.8%) 14326 (� 2.3%)

GFRP humidity exposureUnexposed (control) 36.09 (� 2.9%) 955 (� 6.5%) 26185 (� 7.3%)50 days 35.82 (� 3.4%) 862 (� 4.8%) 23813 (� 6.0%)100 days 36.31 (� 2.6%) 857 (� 5.9%) 26116 (� 6.5%)180 days 35.71 (� 1.8%) 835 (� 5.4%) 27664 (� 7.8%)

SOR6 12/19/98, 3:36 PM172


described earlier. Mean data points are quoted, the values in parenthesesrepresenting the maximum deviation as a percentage. The moduli of theCFRP specimens subjected to thermal cycling exhibited a consistent in-creasing trend with increasing exposure time, whereas those values for theGFRP specimens showed a decreasing trend. The lack of consistent trendsfor all other values suggests that the exposure conditions had no conclusiveinfluence on those material properties.

All tests were conducted at a time of 24h after removal from the tem-perature cabinet to enable them to reach the surface dry state; the speci-mens were stored under ambient laboratory conditions during this time.

6.6.3 Key observations

The observations that the ROBUST consortium made from these environ-mental tests were that:

• The adhesive mechanical properties were increased by the temperaturecycling, the modulus of elasticity increasing rapidly initially. Thepostcuring effect of temperatures greater than the cure temperature wasbelieved to be responsible for the improvement in properties.

• The CFRP composite experienced an increase in modulus and strengthdue to the thermal cycling between �20 °C and �50 °C. This was possi-bly caused by postcuring of the epoxy matrix, indicating the superiorperformance of CFRP compared with GFRP which experienced nosignificant improvement.

• Under elevated temperature and humidity there is softening of theadhesive, whilst the CFRP composite became stiffer and stronger but toa lesser extent than under thermal cycling exposure. These effects werenot considered to be significant.

6.7 Influence of surface treatment and effects of

environment on joints and interfaces

6.7.1 Introduction

It was stated in Section 6.3 that optimisation of substrate surface conditionsand pretreatments often represents the key to maximising joint durability.Whilst the surfaces of both concrete and polymer composite materials arerelatively stable, they do require adequate preparation for structural bond-ing (Section 3.4). Central to optimisation of surface conditions to providejoints of high strength and durability is the adoption of appropriate me-chanical test techniques and methods.

There is a clear emphasis at the beginning of this section on the use of testprocedures which seek to test the integrity of bonded interfaces. Later

SOR6 12/19/98, 3:36 PM173


subsections deal with an assessment of the overall performance of smallscale beams.

6.7.2 Joints involving adhesive–concrete interfaces

The tensile pull-off test is probably the most frequently used on-site testmethod for assessing the quality of concrete and also for examining theadhesion of repair materials and coatings. Tests may be conducted todetermine both initial and long term ‘strength’ following environmentalexposure.

The partially cored pull-off test (PrEN 1542, 1996) was employed in theROBUST project. This was used to assess the effects of concrete surfacepreparation and concrete surface moisture content on the bond perform-ance of Sikadur 31PBA, a two-part cold cure paste epoxy resin. Essentially,steel dollies were bonded to the concrete pull-off locations as defined by50mm diameter partial cores to a depth of 15mm (Fig. 6.3b). The bondlinethickness was of the order of 2mm and a minimum time period of 7 dayselapsed before any pull-off tests were carried out. The characteristic con-crete cube strength was 42MPa and the typical age of the concrete at thetime of testing was around 20 weeks.

The test parameters used were:

• Three concrete surface moisture contents, generated by:– 14 weeks laboratory ageing (20 °C/50% rh)– 8 weeks laboratory ageing, 3 weeks water immersion at 20 °C, fol-

lowed by 3 weeks laboratory ageing– 11 weeks laboratory ageing, followed by 3 weeks water immersion at

20 °C.• Two types of surface treatment:

– low pressure alumina grit blasting (just exposing small aggregateparticles)

– high pressure chilled iron grit blasting (exposing medium sized ag-gregate particles).

• Two types of ageing following exposure:– 50 24 h freeze–thaw cycles (�18 °C to �18 °C)– laboratory ageing (20°C/50% rh)

The results of the concrete pull-off tests are given in Table 6.3. Theaverage pull-off strengths ranged from 1.65–2.91MPa with the lower figurescorresponding to concrete subjected only to laboratory conditions prior toadhesive bonding. Curiously, the highest figures corresponded to jointsmade with the dampest concrete and then subjected to freeze–thaw cycling.In all cases the locus of failure was cohesive within the concrete. No distinc-tion could be drawn between the relative merits of ‘low’ and ‘high’ pressure

SOR6 12/19/98, 3:36 PM174

Environm

ental durability175

Table 6.3 Results of partially cored pull-off tests (Rahimi, 1996)

Concrete conditioning Concrete surface Conditioning after bonding Test results Failurebefore bonding treatment mode

low high 50 24 h ambient average averagepressure pressure freeze–thaw laboratory failure load pull-offalumina chilled iron cycles (�18 °C conditions (kN) with SD strengthgrit blasting grit blasting to 18 °C) (20°C, 50% rh) in brackets (MPa)

within14 weeks under ambient X X 5.08 (1.13) 2.59concrete

laboratory conditions � X X 3.24 (1.2) 1.65 �(20 °C, 50% rh) X X 5.20 (0.75) 2.65 �

X X 3.80 (0.34) 1.94 �

8 weeks under ambientlaboratory conditions plus

3 weeks water immersionX X 5.26 (0.7) 2.68 �

at 20 °C followed by � X X 4.38 (1.41) 2.23 �

3 weeks under ambientX X 5.62 (1.36) 2.86 �

laboratory conditionsX X 4.54 (0.45) 2.31 �

11 weeks under ambient X X 5.36 (0.88) 2.73 �laboratory conditions X X 5.72 (0.69) 2.91 �followed by X X 5.14 (0.93) 2.62 �

3 weeks water immersion � X X 4.78 (1.46) 2.43 �at 20 °C

Failure loads represent the average of four to five test results.

175

SO

R6

12/19/98, 3:36 PM

175


grit blasting. Furthermore the dampness of the substrate did not affect thebond between this particular adhesive and the concrete, at least as deter-mined by this test procedure.

6.7.3 Joints involving adhesive–FRP interfaces

The lap shear joint is that used almost universally in testing adhesives orsurface preparation techniques. It owes its popularity to its convenience ofmanufacture and test, as well as to the fact that the adhesive is subjected tocleavage as well as to shear. It thus simulates the actual use of an adhesivein a variety of applications, including that in plate bonding.

The single lap shear joint is suited to assessing qualitatively the adhesionand bond integrity between materials, since joints made with relatively stiffadhesives and thin adherends fail by a cleavage mechanism. Data so gener-ated are qualitative only, but the locus of joint failure can provide informa-tion on the durability of the bond. Joints may be subjected to environmentalexposure both unstressed and stressed in suitable fixtures.

Single lap shear joints fabricated generally in accordance with ASTMD3163 (1973) were used in the ROBUST project. In fact the adherendcoupons measured only 60mm � 20mm, the joint overlap was set at 10 mm,and the bondline thickness of Sikadur 31PBA adhesive was 0.5mm. Thejoints were postcured at 50°C for 6h prior to environmental exposure. Thetest parameters used were:

• peel-ply surface treatments for the adherends• two types of polymer composite adherends (fabricated from prepreg

materials): GFRP (1.8 mm thick) and CFRP (1.2mm thick)• unstressed exposure for up to 1.5 years at 40 °C/95% rh• three levels of exposure at 40 °C/95% rh under stress at 10% to 30% of

the initial control joint strength.

A comparison of the results of unstressed exposure is shown in Fig. 6.4(Rahimi, 1996). The somewhat erratic response in the first few weeks sim-ply indicates some plasticisation of the bondline perimeter, resulting in arelief of bondline stress concentration and changes in joint behaviour.Thereafter the dominant effect of plasticisation of both the bondline andthe adherends takes over, resulting in an overall reduction of perhaps 15%in the joint strength. The reduction in strength stops somewhere between 20and 50 weeks. In all cases the locus of failure remained cohesive within theadhesive layer, except for the initial joints tested. This is very encouragingand indicates a stable adherend/adhesive system.

The results of the stressed joints are shown in Fig. 6.5 (Rahimi, 1996).It is clear that as the applied stress increased, the time to failure de-creased, as expected. Again the locus of failure for all joints remained

SOR6 12/19/98, 3:36 PM176


Figure 6.4 Effect of exposure at 40°C/95% rh on the performance ofunstressed single lap shear joints made with CFRP and GFRPadherends.

Figure 6.5 Effect of exposure at 40°C/95% rh on the performance ofsingle lap shear joints made with CFRP and GFRP adherends, stressedat different levels.

SOR6 12/19/98, 3:36 PM177


within the adhesive layer. The data suggest that joints made with suchsmall bonded areas are unable to carry sustained loads for long periods, butthis should not cause alarm for the joint size-effect reasons stated inSection 6.5.3.

For the case of the particular adherend/adhesive systems studied here,the results indicate that there was some water-induced plasticisation ofthe small scale laboratory joints. This was caused predominantly by somesoftening of the adhesive layer, an outcome which small joints are far moresensitive to than large ones, but which nevertheless can be designed for.The most important point, which gives great confidence, is that the bondbetween the adhesive and FRP materials was stable under hot/wetconditions.

6.7.4 Effects of environment on strengthened smallscale beams

To investigate the environmental effects on the ROBUST plated beamsystem, RC beams 0.8m long and 100mm � 100mm cross-section wereused. The reinforcement of the concrete consisted of three, 6mm diametersteel rebars in the tensile region and two, 6mm diameter rebars in thecompression region; the beams were over-reinforced in shear. The CFRPcomposite plates had a cross-section of 65mm � 0.7 mm. The concrete was54MPa grade with an elastic modulus of 35GPa. The beams were exposedto the same environmental history as described for the individual compo-nents in Section 6.6.2; the thermally cycled specimens were unloaded duringthe exposure but the humidity-exposed specimens were loaded in four pointbending as specified below.

The testing regime for the post-environmental exposure (thermal cy-cling) utilised a four point loading system in which the constant momentlength was 100mm and the span was 740 mm. This arrangement was alsoused during the humidity exposure and the total load applied to the beamduring this time was 17kN. This represented the value of load that wouldcause yield of the steel reinforcement and this was well above the service-ability load of that beam.

Table 6.4 shows the results for the plated RC beams after a period of 180thermal cycles between �20 °C and �50 °C. The values obtained for thebeams in terms of comparative strengthening and stiffening effects areshown relative to identical beams tested directly to failure.

The beams exposed to the hot humid conditions showed no appreciablecreep (measured by removable deformation dial gauge) or fall-off of load(measured by load cells permanently in position). There was no detectabledegradation of either of the beams.

SOR6 12/19/98, 3:36 PM178


Table 6.4 Strengthening comparison of 0.8 m long durability beams subjectedto thermal cycling (Garden and Hollaway, 1997)

Beam Yield load Ultimate load Postyield stiffness(kN) (kN) (kNmm�1)

Unexposed (control) 24.0 48.9 7.3Beam 1 26.0 46.9 5.4Beam 2 25.0 44.0 6.3

6.8 Other factors affecting service performance

6.8.1 Freeze–thaw action

Consideration has been given to repairing bridges in cold countries, particu-larly where structural strength needs to be restored following damage bycorrosion to reinforcing steel. Green and Soudki (1997) devised some ex-periments on small columns and beams to investigate the effects of freeze–thaw action. One 24 h freeze–thaw cycle consisted of 16h at �18 °C,followed by 8 h in water at �18 °C. Cylinders 150 � 300mm wrapped withCFRP sheets were subjected to 200 cycles; beams 1200 � 150 � 100 mmwere plated with both CFRP and GFRP, and then subjected to 50 cycles.They concluded that the FRP sheets were very effective at restoring thestrength of the columns damaged by freeze–thaw action and that thestrengthened beams behaved in a similar way to their counterparts main-tained at laboratory ambient temperatures. The FRP materials and thebond between the concrete and the FRP sheets were unaffected by thefreeze–thaw cycling.

6.8.2 Fire

The vinylester polymer (Palatal A430-01), which was the polymer used informing the pultruded composite plate in the ROBUST project, is a non-fire-retardant resin. During the project no attempt was made to increase thefire resistance of the pultruded composite as the aim of the research workwas specifically to investigate strengthening of the reinforced concretebeams by composite plate bonding.

However, polymer materials composed of carbon, hydrogen and nitro-gen atoms (i.e. they are organic materials), are all inflammable to varyingdegrees and could suffer some deterioration if exposed to fire unless someform of protection against it is provided (Hollaway, 1993). It is possible toincorporate additives into the resin formulations or to alter their structure,thereby modifying the burning behaviour and producing a composite with

SOR6 12/19/98, 3:36 PM179


enhanced fire resistance. Furthermore, it is possible to provide a fire protec-tive coating after the plate is bonded into position onto the external surfaceof the structural unit.

The fire hazards of polyester resins (vinylester is an enhanced polyester)may be reduced by incorporating halogens; halogens are one of the fluorine,chlorine, bromine and iodine family of chemicals. Fire hazards can also bereduced by combining synergists such as antimony oxide into the resinformulations; these are commonly known as hot acid resins. The common-est way, and indeed the cheapest, of obtaining these is by the addition ofchlorinated paraffin and antimony oxide or by the utilisation of halogenatedphosphates such as trichloroethylphosphate. Such systems when combinedwith fibres would conform to the British Standard Test for Fire Retardance,Class 2, of BS 476 Part 7.

To enable polyester composites to be used in building applications, resinshave been developed which can attain a Class 1 rating, thus complyingwith the Building Regulations. Thus numerous other means of intro-ducing halogens, either by additives such as pentabromotoluene andtris(dibromopropyl)phosphate or by reactants such as tetrabromophthalicanhydride or dibromoneopentyl glycol have been developed. The effects ofalternative synergists and inert fillers have resulted in the availability ofclear and opaque resins which can be fabricated into composites to give aClass 1 fire rating.

An alternative method of fire protection is by employing intumescentcoatings. These can be based on polymer resins with three additives:

• a source of phosphoric acid• a polyhydroxy compound which reacts with phosphoric acid to form a

char• a blowing agent.

In a fire situation these coatings yield a carbonaceous expanded charwhich protects the underlying composite. Structures treated by this methodcan develop fire ratings of Class 1 and Class 0.

Deuring (1994), undertook fire tests on RC beams, plated with carbonfibre/epoxy matrix polymer composites, bonded with Sikadur 30 S-02 andSikadur 31 SBA S-08 epoxy adhesive. Some of the plates were protectedwith 60mm thick PROMATECT-L.

When decisions are required regarding fire protection, it is advisable forthe engineer to discuss the most appropriate system to be used with theresin manufacturers. If fire protection is required for plate bonding it isunlikely that the additives, which would be incorporated into the resin,would affect the strength of the composite significantly, but the in-serviceproperties of colouring and UV might be affected. However, as the polymercomposite plate is located on the soffit of a concrete beam, and therefore

SOR6 12/19/98, 3:36 PM180


protected from the direct rays of the sun, any degradation from this sourcewill be minimal. Additionally, any change in the colouring of the compositewill be masked by the black carbon fibre; consequently, from the weather-ing point of view, degradation will be insignificant. It is worth mentioning,however, that if UV degradation is a problem, UV stabilisers can be addedto the liquid resin in addition to any additives for fire resistance.

6.9 Summary

The main considerations relating to the environmental durability of exter-nally bonded reinforcement are the quality of the materials used and theintegrity of the adhesive bonds. The ROBUST system of strengtheningrepresents an environmentally stable system with good durability as dem-onstrated in laboratory experiments. This behaviour results from the use ofgood quality unidirectional CFRP/vinylester matrix material manufacturedby the pultrusion process, an epoxy adhesive which exhibits good all roundphysical and mechanical characteristics (together with low permeability towater), excellent adhesion to concrete and excellent adhesion to CFRPconferred by use of the peel-ply surface preparation technique coupledwith the wetting characteristics of the adhesive. It has been shown that thematerials, and bonds to them, are very stable within the normal range ofoperating conditions anticipated.

6.10 References

ASTM D3163 (1973) Standard Test Method for Determining the Strength ofAdhesively Bonded Rigid Plastic Lap-shear Joints in Shear by Tension Loading,American Society for Testing and Materials, Philadelphia.

Bodnar M J (ed.) (1977) Durability of Adhesive Bonded Structures, J AppliedPolymer Science, Applied Polymer Symposia 32, New York, J Wiley & Sons.

Brewis D M, Comyn J and Shalash R J A (1982) ‘The effect of moisture andtemperatures on the properties of an epoxide-polyamide adhesive in relationto its performances in single lap joints’, Int. J. Adhesion Adhesives 2(4) 215–222.

Brockmann W (1983) in Aspects of Polymeric Coatings, ed. K.L. Mittal, PlenumPress, New York, pp 265–280.

BS EN 29142 (1993) Adhesives – Guide to the Selection of Standard LaboratoryAgeing Conditions for Testing Bonded Joints, BSI, London.

Clarke J L (ed.) (1996) ‘Structural design of polymer composites’, EUROCOMPDesign Code and Handbook, London, E & F N Spon.

Comyn J (ed.) (1985) Polymer Permeability, Elsevier Applied Science, London.Comyn J (1983) ‘Kinetics and mechanisms of environmental attack’, in Durability of

Structural Adhesives, ed. A J Kinloch, London, Applied Science, Chap. 3.Deuring M (1994) Brandversuche an nachtraglich verstärken Trägern aus Beton,

Report No. 148795, EMPA (in German).

SOR6 12/19/98, 3:36 PM181


Garden H and Hollaway L C (1997) Environmental Durability of Composites,Internal report of the University of Surrey.

Green M F and Soudki K A (1997) ‘FRP strengthened concrete structures in coldregions’, Proc. Int. Conf. on Recent Advances in Bridge Engineering – AdvancedRehabilitation, eds. U Meier and R Betti, EMPA Dubendorf, Switzerland, pp219–226.

Hollaway L C (1993) Polymer Composites for Civil and Structural Engineering,Blackie Academic and Professional, Glasgow.

Hutchinson A R (1986) Durability of Structural Adhesive Joints, PhD Thesis, Uni-versity of Dundee.

Kinloch A J (1982) J. Mater Sci 17 617–651.Kinloch A J (1983) Durability of Structural Adhesives, Applied Science, London.Kinloch A J (1987) Adhesion and Adhesives: Science and Technology, Chapman and

Hall, London.Kinloch A J and Shaw S J (1981) in Developments in Adhesives – 2, ed. A J Kinloch,

Applied Science, London, pp 83–124.Loos A C, Springer G S, Saunders B A and Tungl R W (1981) ‘Moisture absorp-

tion of polyester E-glass composites’ in Environmental Effects on CompositeMaterials, ed. G S Springer, Technomic, Westport, USA, pp 1–62.

Mays G C and Hutchinson A R (1992) Adhesives in Civil Engineering, CambridgeUniversity Press.

Minford J D (1983) in Durability of Structural Adhesives, ed. A J Kinloch, ElsevierApplied Science, London, pp 135–214.

Norwood S R (1994) ‘Fibre reinforced polymers’, in Handbook of Polymer Com-posites for Engineers, ed. L C Hollaway, Woodhead Publishing, Cambridge.

PREN 1542 (1996) Products and Systems for the Protection and Repair of ConcreteStructures – Test Methods – Pull-off test, BSI, London.

Rahimi H (1996) Strengthening of Concrete Structures with Externally Bonded FibreReinforced Plastics, PhD Thesis, Oxford Brookes University.

Ripling E J, Mostovoy S and Bersch C F (1971) ‘Stress-corrosion cracking ofadhesive joints’, J. Adhesion 3(2) 145.

Shen C H and Springer G S (1981) Effects of temperature and moisture on thetensile strength of composite materials’, in Environmental Effects on CompositeMaterials, ed. G S Springer, Technomic, Westport, USA, pp 79–92.

Tu L and Kruger D (1996) ‘Engineering properties of epoxy resins used as concreteadhesives’, ACI Mater J 93(1) 26–35.

Venables J D (1984) ‘Review: durability of metal-polymer bonds’, J Mater Sci 192431–2453.

SOR6 12/19/98, 3:36 PM182

Time-dependent behaviour and fatigue 183

183

7Time-dependent behaviour and fatigue

R B A R N E S A N D H N G A R D E N

7.1 Introduction

This chapter is divided into two parts, the first concentrating upon creepdeformation characteristics of the component parts of a reinforced concrete(RC) plated beam and the beam system itself. The second part of thechapter discusses the fatigue characteristics of the component parts and theplated beam. The two parts will present results of tests conducted duringthe ROBUST project to determine these two properties for the ROBUSTplating technique.

Part A Time-dependent behaviour

7.2 Introduction

The resin systems used to manufacture both fibrous composites and theadhesives subsequently used to bond them are polymeric materials and, assuch, the same considerations in terms of time dependency apply to both. Inany situation where an adhesive connection is required to transmit stressesdue to sustained load, the possibility of creep in the adhesive must beconsidered. Long term bond integrity implies, amongst other things, theavoidance of excessive creep and the possibility of creep-induced failuresthroughout the service life.

In this particular application, since the adhesive is the component whichtransfers load to the external plate, creep within the bondline may affect thestructural performance of the strengthened member. The action of creepon an adhesive is to cause a degradation of the effective modulus of thematerial, or a loss of rigidity. This may reduce the stress transferring abilityof the adhesive and thus the efficiency of the strengthening system. Simi-larly, creep of the external fibre reinforced polymer (FRP) plate itselfwould result in time-dependent increases in member deflection under the

SOR7 12/19/98, 3:38 PM183


action of a constant load. The fact that three of the materials involved in theapplication are viscoelastic and that the behaviour and interaction of thesematerials may affect the structural performance of the overall system, makeit necessary to understand the time dependency of this system.

For applications involving bridge rehabilitation or upgrading, the exter-nal plate is most likely to be bonded to the structure without additionalpropping, so that none of the member’s self weight is supported by theplate, the system only providing strengthening for superimposed or liveloads applied subsequently. In such circumstances, long periods of sus-tained loads are unlikely. However, if the technique of pretensioning theexternal FRP plate prior to bonding is implemented, the plate will remainheavily loaded at all times and the adhesive will be under sustained stress.Furthermore, for building strengthening applications, the level of live loador superimposed dead load may be reasonably constant for long periods oftime, placing the plate, and thus the adhesive, under conditions of sustainedstress.

7.3 Time-dependent characteristics of concrete

The relationship between stress and strain for concrete is a function of time.Under sustained stress, concrete undergoes a gradual increase in strain,referred to as creep, most conveniently taken as an increase in strain abovethe initial elastic value. Creep generally has little effect on the strength of astructure, but will cause an increase in deflections under service loads and aredistribution of stress. Considerable research effort has been put into thesubject (for example, Evans and Kong, 1996; Neville, 1970; ACI, 1982) anda complete discussion will not be given here.

7.4 Time-dependent characteristics of steel

When metal is subjected to a stress, slow plastic deformation, usuallyreferred to as ‘creep’, can occur under constant load, even when the stresslevel lies below the elastic limit. Creep associated with reinforced and pre-stressed concrete can have a considerable effect upon the durability of thematerial. The satisfactory performance of the rebars depends upon theirability to maintain a high elastic tensile stress throughout the lifetime of thestructure. Further information on the creep of steel in conjunction withconcrete can be found in Neville et al. (1983).

7.5 Time-dependent characteristics of adhesives

The mechanical properties of polymers have characteristics of both elasticsolids and viscous fluids, and hence they are classified as viscoelastic mate-

SOR7 12/19/98, 3:38 PM184


rials. On application of load, the material deforms instantaneously like anelastic solid. However, if the load is maintained, the material continues todeform to some extent in a manner similar to a viscous fluid. Therefore, forfull characterisation of such a material, a knowledge of its time-dependentresponse is necessary. The transition from relatively elastic behaviour torelatively viscous behaviour often occurs within a timescale in the range ofconcern for practical application.

The viscoelastic behaviour of polymers is well documented; a compre-hensive text is given by Ferry (1980). Viscoelastic effects can result in creepdeformation or stress relaxation. Figures 7.1 and 7.2 illustrate a simplifiedcreep strain–time relationship for a polymer under a uniaxial stress stateand for a shear creep strain–time relationship for an adhesive lap joint,respectively. Creep may lead to delayed fracture, termed creep rupture,under long term loading. In this case, the applied stress should be lowerthan that required to cause fracture under monotonic loading conditions.The time-dependent strain of the material divided by the constant appliedstress is termed the creep compliance. Since adhesives are polymeric

Figure 7.1 Simplified creep strain–time relationship for a polymerunder uniaxial stress state.

SOR7 12/19/98, 3:38 PM185


Figure 7.2 Schematic representation of a creep versus time graph fora structural adhesive joint.

materials, they exhibit viscoelasticity and as such are subject to time-dependent behaviour.

The creep of adhesives in general is affected by a number of parameters,including the chemical nature of the resin system, the volume and types offiller used, the dimensions of the bond area and the absorption of plasticis-ers such as moisture or oil. However, the main factors are the level of theapplied load, time and temperature. In addition, the magnitude of the creepresponse for thermosetting resin systems is highly dependent on its curehistory, particularly if under working load the resin is continuously loadedand is also at an elevated temperature. Creep rate varies with stress level,and in general the higher the stress the greater the creep rate. At lowapplied stress levels, many polymers behave in a viscoelastic manner, inwhich the strain response of the material at any time of loading is a linearfunction of the constant applied stress. This implies that the principle ofsuperposition can be used for changes in applied stress (Hollaway, 1993).However, at higher stress levels, the viscoelastic behaviour of polymers canbecome highly non-linear. Perhaps of greatest importance to structuraladhesive joints in civil engineering is behaviour at relatively low stresslevels, since bonded joints tend to be designed to withstand low stresses fora long period of time rather than the higher stresses which may cause rapidcreep to failure. For example, in steel plate bonding applications, it isrecommended that any sustained stress in the adhesive bonded joint bekept below 25% of the short term joint strength to minimise creep effects(Mays, 1993).

SOR7 12/19/98, 3:39 PM186


When an adhesive is placed under load, the stress is not distributedevenly throughout the adhesive but concentrates in certain areas within thebondline. With lap shear joints for instance, the maximum stresses occur atthe ends of the overlap. It is these maximum stresses which are thought todetermine the overall creep behaviour of the joint. Since typical jointsinvolve high stresses near the free ends, even relatively low load levels mayresult in significantly non-linear behaviour. As such, the practical approachboth to delaying the onset of creep and to reducing its rate is by providinglarge bonded areas.

The ambient temperature also affects the occurrence and rate of creepdeformation. As the temperature passes the glass transition temperature,Tg, of an adhesive, there is a marked change in the creep properties exhib-ited by an adhesive bonded joint (Adams and Wake, 1984). For the major-ity of adhesives, an increase above the normal operating temperatures forwhich the material is designed causes an increase in the creep rate; thenonce the Tg is reached, a further increase occurs. The creep behaviour of anadhesive at elevated temperatures can be affected by a number of factorsincluding the proportion of fillers, the shape of the specimens, the stresslevels encountered and the extent of the temperature rise. To limit theeffects of creep under sustained load, an adhesive possessing a Tg wellabove the service temperature is required. In general, an increase in relativehumidity plasticises the adhesive in the joint, reducing its stiffness andstrength. It has also been found that water or high humidity has the effectof lowering the Tg of the material. If the range through which the Tg dropscoincides with the temperature of the joint then this will greatly affect theadhesive strength and cause an increase in creep. However, if the rangethrough which the Tg drops is far from the operating temperature of thejoint then the effect of moisture on the modulus of elasticity of the adhesivewill be of primary importance, since this would also affect the creep prop-erties of the adhesive. In general, the more highly cross-linked the hardenedadhesive structure and the higher the curing temperature and hence Tg, thebetter the creep resistance (Mays and Hutchinson, 1992).

7.5.1 Creep testing of structural adhesives

There are basically two different types of creep test available for the char-acterisation of structural adhesives; these involve the utilisation of bulkmaterial tested in tension and the adhesively bonded joints using a single ordouble lap joint. The latter test more nearly simulates the material as it isused in practice.

The creep of Sikadur 31, a two-part epoxy resin adhesive, similar to thatused in the ROBUST project, was studied at varying stress and temperaturelevels alongside a toughened single part hot cured epoxy paste (Permabond

SOR7 12/19/98, 3:39 PM187


ESP 110) which was examined under identical long term loading (Larkand Mays, 1984; Mays, 1990). A double lap steel joint was employed, theadherends being manufactured from 25.4mm � 6.4mm mild steel with anoverlap length of 80mm. The two-part adhesive was cured at room tem-perature whilst the hot cured adhesive was cured at 100°C. Control testswere performed at room temperature at an age of 7 days, the failure loadbeing recorded and the mean shear strength determined by dividing thefailure load by the bonded area.

Creep testing rigs based on the dead load and lever principle, with acapacity of up to 45kN, were employed. Room temperature (20°C) andrelative humidity (50% rh) were maintained by an air conditioning system.Where specimens were tested at elevated temperatures of up to 65 °C,the temperature was maintained using a heating tape coiled around theadhesive joint.

Room temperature creep curves for the above two-part epoxy resinadhesive, expressed in terms of log strain against log time (h), are presentedin Fig. 7.3. The mean shear stress at which each test was conducted is shownadjacent to each curve. It can be seen that the initial slope of the logarithmiccreep curve increases with applied stress. It should be mentioned that thestress values used in these tests are much greater than would occur inpractice.

The relationship between the applied mean shear stress and time tofailure is illustrated in Fig. 7.4 for the two-part cold cure adhesive. Examin-ing Fig. 7.4 reveals a small reduction in performance between tests at 20°Cand 40 °C but a more significant reduction at 55 °C. Figure 7.5 shows theshear stress versus time to failure for the single-part hot cure epoxy resinadhesive and it will be seen that there is a small reduction in performanceat 55 °C and 65 °C compared with 20 °C. The glass transition temperaturesfor the two adhesives were measured at 44°C and 105°C for the cold cureand hot cure adhesives, respectively. The poor creep performance of thecold cure adhesive at 55 °C (a value above its Tg) demonstrates the effect oftemperatures above the glass transition value.

From a civil engineering point of view, it is important to establish themaximum values of stress which can be sustained so as to ensure survival forthe design life of the structure. In Table 7.1 (Barnes and Mays, 1988) thismaximum stress value (expressed as a fraction of the short term strength) isdisplayed for service times of 1 year, 30 years and 120 years. The valueswere obtained by extrapolating the logarithmic stress versus time to failuremean lines, a method which must be undertaken with some caution. Basedon the statistical analysis applied to the results a lower bound was alsoobtained.

This work indicates that the sustained stress in an adhesive joint shouldnot exceed 25% of the short term strength for the normal design life of the

SOR7 12/19/98, 3:39 PM188

Time-dependent behaviour and fatigue

189

Figure 7.3 Creep curve for two-part epoxy resin adhesive at 20°C (Mays, 1990).

189

SO

R7

12/19/98, 3:39 PM

189

190S


Figure 7.4 Shear stress versus time to failure for two-part cold cure epoxy resin adhesive (Mays, 1990).

SO

R7

12/19/98, 3:39 PM

190


191Figure 7.5 Shear stress versus time to failure for single-part hot cure epoxy resin adhesive (Mays, 1990).

SO

R7

12/19/98, 3:39 PM

191


Table 7.1 Relationship between normalised stress and time to failure

Adhesive Temperature (°C) Normalised stress at time to failure of

1 year 30 years 120 years

Cold cure 20 0.69 0.63 0.6Cold cure 40 0.46 0.38 0.36Cold cure 55 0.18 0.11 0.09Hot cure 20 0.67 0.63 0.61Hot cure 55 & 65 0.41 0.34 0.32

structure. Also, due to the deleterious effects of temperature, adhesivebonded joints should not be exposed to service temperatures in excess ofthe glass transition temperature for that adhesive.

7.6 Time-dependent characteristics of plated beams

using steel plates

Most studies on the time-dependent behaviour of reinforced concretebeams strengthened with externally bonded steel plates have combined theeffects of weathering and sustained loading. However, the durability ofadhesive bonds is discussed in Chapter 6, consequently this section willfocus on the effects of sustained load.

Since 1972, research on plate bonding has been conducted at the SwissFederal Laboratories for Materials Testing and Research (EMPA) atDubendorf. The early work (Ladner and Weder, 1981) included studies ofthe long term performance of steel plated beams under sustained load.

Sixty-six reinforced concrete beams (150mm � 250mm � 2400mm)were strengthened with external steel plates (3mm � 100mm � 1950mm)and prior to strengthening the beams were loaded to produce flexuralcracking. Forty-one beams were loaded with a constant load of value 28%of the failure load, some of the beams also being subjected to weatheringin outdoor test areas. The remaining beams were used as controls, six ofwhich were statically tested before the onset of the creep programme.There was little difference in failure load or failure deflection between thesebeams and beams which had been subjected to sustained loading for oneyear prior to static testing. Although the majority of creep will have takenplace during this period, the length of time for this creep investigation wasshort.

The increase in central deflection with time was also monitored withan average creep of 1.3mm over 800 days, the initial elastic deforma-tion upon application of the load averaging 2.3mm. Creep measurements

SOR7 12/19/98, 3:39 PM192


over 10 years on another plated reinforced concrete beam produced thesame trend in the results expected of a conventional reinforced concretebeam.

Concerns have been expressed by the Institution of Civil, Municipal andStructural Engineers Standing Committee on Structural Safety (StandingCommittee on Structural Safety, 1983) about the long term performance ofresins. Consequently, the Transport and Road Research Laboratory, UK(TRRL) undertook a programme of testing to determine the long termproperties of steel plated reinforced concrete beams (Calder, 1990). Creeptests on plated unreinforced concrete beams showed no statistically signifi-cant differences in failure loads between loaded specimens and unloadedcontrols following 10 years’ exposure. This would indicate that if creep ofthe adhesive does take place the failure of the beam system is not signifi-cantly reduced.

Reinforced concrete beams were also used in the TRRL work, half of thebeams being precracked prior to plating. All the beams were then heldunder sustained load sufficient to produce and hold open flexural cracks.Control beams were kept in a laboratory environment (20 °C, 65% rh)whilst the remaining specimens were exposed at three outdoor exposuresites, chosen to represent conditions encountered throughout the UK. Twodifferent resins were used, resin I (Ciba Geigy XD800) and resin II(Colebrand CXL83), both of which were structural epoxy adhesives. Beamsplated with resin I showed no difference in failure load following eightyears’ sustained load when compared to control beams tested soon afterplating. Resin II, however, was more susceptible to environmental attack;those beams which survived eight years’ exposure failed prematurely due toplate debonding. The tests showed that moisture penetration leading tosteel corrosion was a far more serious threat to long term performance thancreep under sustained load.

Another UK study of the long term behaviour of plated RC beams wasconducted in an area of high industrial pollution near Sheffield (Swamyet al., 1995). Eight plated beams were maintained under constant load andthe failure loads after 11–12 years’ exposure were found to be in excess ofthe 28 day controls. The increased concrete maturity with time obviouslyhad an influence on these results. There was no significant difference infailure loads in 11–12 year old beams between those under sustained loadand those left unloaded.

In Spain, creep tests with temperatures ranging from �11–�42 °C wereconducted using three different adhesive types and varying adhesive layerthickness (Canovas, 1990). Satisfactory behaviour for three years was re-ported for plated beams with minimum adhesive thicknesses (up to 1mm).As the adhesive layer thickness increased, however, so did creep, as evi-denced by increased central deflections. With one of the adhesives this led

SOR7 12/19/98, 3:39 PM193


to failure of the beam after 420 days. This adhesive was deemed to beunsuitable for future use in plate bonding.

It is generally accepted that the long term performance under load ofsteel plated reinforced concrete beams is satisfactory provided that a suit-able adhesive and adhesive layer thickness are utilised.

7.7 Time-dependent characteristics of FRP

component materials and FRP composites

The matrix material of FRP composites is polymeric, and hence thesematerials are subject to viscoelastic behaviour. The creep response of thecomposite material is a combination of elastic fibres which do not creep anda flexible viscoelastic polymer matrix which does. The creep of compositematerials is highly dependent upon:

• the type of polymer and its stress history• the direction of alignment, type and volume fraction of fibre

reinforcement• the nature of the applied loading• the temperature and moisture conditions to which the element is

exposed.

A number of investigators have reported on the time-dependent natureof polymer composites in general, for example Wu and Ruhmaan (1975),Crossman and Flaggs (1978) and Schaffer and Adams (1981). Creep ofthese materials can be substantial, particularly when subjected to highloads, elevated temperatures or moisture absorption. Scott et al. (1995)present a review of literature concerned with the creep behaviour of FRPcomposites, with particular reference to the development of acceleratedtest methods for predicting long term performance of FRP for structuralhighway applications.

7.7.1 Mechanisms of FRP creep

There are two distinct mechanisms by which FRP materials can creep, theoccurrence of which depends on the state of the microstructure of thematerial. In predicting the creep response of a composite material, it isgenerally assumed that the material microstructure remains intact, with thecomposite structure of the fibre and matrix deforming as one unit, the resinphase transmitting stress between the fibres. Consequently, the creep of thecomposite will be due solely to the viscoelastic flow of the polymer matrix.This assumption may be valid at elevated temperatures near to the Tg of thepolymer, but in the general working temperature range for constructionmaterials the primary mechanism of creep is more complex. It is likely that

SOR7 12/19/98, 3:39 PM194


the majority of the observed creep, stress rupture and stress relaxation atambient temperatures is due to the time-dependent growth of fibre–matrixdebonds and resin cracks within the material microstructure (Johnson,1979).

It is generally accepted that microdamage in the form of cracks is initi-ated when a component of force acts perpendicular to the fibre axis. Cracksmay also be initiated due to differences in the thermal expansion coeffi-cients of the constituent materials (Sillwood, 1982). The high density ofgeometric discontinuities in fibrous composites, such as fibre ends andvoids, and the brittle nature of the resin systems used, also aids the initiationof microdamage. This damage growth mechanism can result in non-linear,stress-dependent viscoelastic behaviour over a wide range of applied stress,even when the material is lightly loaded (Howard and Hollaway, 1987).

The occurrence of creep under sustained static loading is therefore acombination of bulk material strain and microflaw initiation, both of thesemechanisms being time dependent due to the viscoelastic nature of thepolymer matrix. However, it is difficult to distinguish from the macroscopicbehaviour of a composite material under load, the relative combination ofthe bulk viscoelastic deformation of the polymer and that of the microflawgrowth. A number of investigators, for example Aboudi (1990), have re-ported sudden jumps in the creep strain which may correspond to the onsetand growth of damage as opposed to the smoother process of the bulk creepstrain of the resin alone.

7.7.2 Creep characteristics of FRP composites

The synthesis of data reported on the creep characteristics of polymercomposite materials is difficult due to the diversity of the resin and fibretypes, fibre orientations and combinations, testing environments and fabri-cation processes. The extent to which the stiffness and strength characteris-tics of a composite material are affected by the time-dependent mechanicalproperties of the matrix can generally be related to the proportion of loadbeing carried by the polymer. Significant time dependence is demonstratedwhen the applied load is shared between the fibre and the matrix, or iscarried mainly by the polymer matrix. The effect is most pronounced inlow fibre volume fraction materials and in tests normal to the predominantfibre orientation. Provided that the contribution of the matrix phase to theoverall stiffness of the material is small, the effect of the time-dependentchanges in the properties of the matrix material on the overall composite isalso likely to be small. Since time dependence is reduced as the proportionof the applied load carried by the reinforcing fibres increases, it follows thatvirtually time-independent behaviour can be achieved through the use of ahigh volume fraction of high stiffness fibres placed unidirectionally in the

SOR7 12/19/98, 3:40 PM195


predominantly loaded direction. Research on the creep of unidirectionallyreinforced laminates is reported by, amongst others, Dillard and Brinson(1983), Hiel et al. (1983), Tuttle and Brinson (1986), Ponsot et al. (1989) andAboudi (1990).

In the context of external plate bonding, the plate performs a simple task,acting as a uniaxially stressed tensile component as the composite memberdeflects. The direction in which the strengthening material will be loaded istherefore known. If a fibre reinforced composite is to be utilised, it can bemanufactured with all of its strength in this direction as a unidirectionallaminate of continuous fibres. The greater the volume of fibres which can beincluded in the composite, the greater will be its stiffness and strength, andhence the greater its efficiency. It follows from the discussion above thatsuch a material should display minimal time dependence, especially if highstiffness fibres are incorporated.

7.8 Time-dependent characteristics of plated beams

using polymer composite plates

Bonded FRP systems can produce a dual viscoelastic problem, the combi-nation of a viscoelastic adherend and a viscoelastic adhesive. In addition,a third viscoelastic component is introduced when one of the adherendsemployed is concrete subjected to shear in the cover concrete andcompressive stresses above the neutral axis. Although many investigationshave been carried out to evaluate the time dependency of concrete, adhe-sives and FRP materials individually, only one recent study by Plevris andTriantafillou (1994) could be found in the literature in which these materialswere employed in combination. In this study, an analytical procedure topredict the time-dependent behaviour of RC beams strengthened with FRPlaminates was presented. These procedures, implemented by computer,were used in a parametric study to assess the effects of the type and areafraction of FRP material on the long term response of internal stresses andstrains and overall curvature of the member. An age-adjusted effectivemodulus of elasticity was used to represent the behaviour of the concrete incompression, while Findley’s model (Findley, 1960) was utilised for thecomposite material.

The addition of CFRP plates reduced the rate of compressive creep in theconcrete of the plated beams compared with the unplated ones; this de-crease in creep rate was greater for plates with larger cross-sectional area.Furthermore, overall curvature and tensile stresses in the internal steelwere reduced by the addition of external FRP plates; the response of thesteel was shown to be time independent. The response of beams strength-ened with either glass fibre reinforced plastic (GFRP) or carbon fibrereinforced plastic (CFRP) were shown to be similar.

SOR7 12/19/98, 3:40 PM196


7.9 Creep tests conducted during the ROBUST

project

7.9.1 Creep testing of CFRP

In the ROBUST programme of work, a unidirectional CFRP prepreg com-posite manufactured by ‘Cytec’ (‘CYCOM 919 HF – 42%-HS-135-460 (P/N02098)) containing 58% weight fraction of Toray ‘T300’ fibres and 42%epoxy resin (the properties of which have been given in Chapter 3) was heldunder sustained tensile loads of various magnitudes at 22 °C, 40 °C and60 °C and a relative humidity of 50%; the material was the same as that usedin a number of the beam tests from which the static behaviour of platedbeams was determined (Garden et al., 1996; Garden et al., 1997; Gardenand Hollaway, 1997) as discussed in Chapter 4. The higher temperatureswere chosen to accelerate the potential creep process arising from theviscoelastic behaviour of the epoxy resin and to simulate possible servicetemperatures. The variation of longitudinal strain with time for the CFRPis shown in Fig. 7.6 for each temperature under the highest applied tensionwhich was nominally 60% of the composite tensile strength. The slightinitial fall in strain at 40 °C and 60 °C is attributed to the slippage of theCFRP from beneath the aluminium end tabs of the coupon specimens, aneffect observed due to the softening of the adhesive at these elevatedtemperatures, continuing until a new equilibrium load was established.While this slippage was undesirable, the settlement of the strain at the newlevel reflects the absence of any creep of the CFRP under the new stresses;the applied load remained constant during the period of settled strain. Theabsence of any apparent creep is consistent with the previous literaturewhich reports excellent CFRP creep resistance with high fibre volumeratios, even when the material is loaded to a high proportion of its tensilestrength.

7.9.2 Creep testing of plated beams using CFRP plates

In the ROBUST programme of sustained load tests, 1.0 m long reinforcedconcrete beams of 100 mm square cross-section were externally strength-ened with the CFRP plates whose creep behaviour is represented in Fig. 7.6,so it was known that any overall creep of the plated beams under loadwould not have been contributed to by the composite material itself. Atypical deflection response of the 1.0m beams, loaded such that the topextreme surface of the concrete and the internal tensile rebars were underthe same stress magnitude in the unplated and plated cases, is shown in Fig.7.7. These identical stress magnitudes were applied to the two beam systemsto ensure that the presence of the plate and the adhesive would be the only

SOR7 12/19/98, 3:40 PM197


Figure 7.6 Variation of tensile strain under sustained load for a typicalCFRP composite (Garden, 1997).

Figure 7.7 Deflection responses of 1.0m CFRP plated beams undersustained load (Garden, 1997).

cause of any difference in the structural behaviour under sustained load.This state of stress was achieved by loading the plated beam to its rebaryield load, and the unplated beam to that load beyond yield which producedthe same maximum compressive stress in the concrete. The rebars wereplastic beyond their yield strain so it was known that the rebar stresses wereequal in both cases; sustained tensile tests of the rebars confirmed that the

SOR7 12/19/98, 3:40 PM198


Figure 7.8 Concrete compressive strain responses of 1.0m CFRPplated beams under sustained load (Garden, 1997).

internal reinforcement did not contribute to the creep of the beams, thesteel strain having remained constant.

The vertical steps in Fig. 7.7 mark the increases in elastic deflectioncorresponding to the restoration of the load which was necessary twice dailydue to creep of the beams. As this load restoration applied an elastic strainto the beam, the total creep deflection at any time ‘t’ was calculated as thetotal increase in deflection under static load from time zero to time ‘t’,minus the sum of these vertical elastic steps. The creep deflection value ofthe unplated beam was 0.33mm and was relatively large, representing 4.7%of the original 7.00 mm elastic deflection, whilst the creep deflection of theplated beam was 0.14 mm and was only 2.7% of the originally smallerdeflection of 5.10mm. The creep of the unplated beam was 3.7% of the totaldeflection at the end of the creep period, higher than the 2.4% in the platedcase, indicating a reduction in the time-dependent deformation due to thebonded plate.

Figure 7.8 shows the relationship between the midspan concretecompressive strains at the top surfaces of the 1.0 m beams and the timeunder continuous loading. The unplated concrete beam crept at a higherrate and had a magnitude of creep greater than the plated concrete beam,reflecting the deflection behaviour. The ratio of unplated to plated concretebeam creep strain was 2.18, similar to the deflection ratio of 2.36, suggestingthe time-dependent deflection of the plated beam was predominantlyflexural and not influenced by any significant vertical shear creep. At the

SOR7 12/19/98, 3:40 PM199


start of the continuous loading period, significant shear cracks were appar-ent in the shear span.

Further ROBUST creep tests were undertaken at the serviceability loadof these beams. The graphs of deflection against time and creep strain of theconcrete against time were of similar shape to those above. The creepdeflection was approximately one-third of the total deflection for both theunplated and plated beams. Furthermore, the compressive concrete creepwas approximately one-third of the total compressive strain under theserviceability load. From strain gauge data taken at various positions alongthe bonded composite plate, whilst the beam was under a sustained load, anincrease in strain was measured which was also one-third of the total tensilestrain of the composite plate. As the stress/strain relationship for CFRPcomposite material is linear to failure and the material does not creep, theincrease in strain of the composite plate was attributed to the deflection ofthe beams due to creep of the concrete. The influence of the adhesive layeron the global creep response was uncertain but, since the concrete and thecomposite plate both experienced the same proportional long term strainincrements with respect to the total strain, it may be said that the adhesiveexperienced these same proportional strain increments (namely, it experi-enced no creep at all during sustained load). Consequently, the viscoelasticproperty of the adhesive appears to present no particular problems whenthe plated bonded beam is under a sustained load at ambient temperature.The fact that the unplated and plated beams experienced similar propor-tional creep suggests the sustained load behaviour of beams is governedby the compressive creep of the concrete and, therefore, not by any time-dependent effects of the adhesive.

Part B Fatigue behaviour

7.10 Introduction

Dynamic fatigue is the failure of a material or assembled component due tothe application of a large number of stress/strain changes, the magnitude ofwhich is lower than those at the static failure load. Fatigue occurs due toirreversible processes that take place when a cyclic load is applied to amaterial, as will be reviewed in the following sections.

The process of fatigue involves progressive and irreversible deteriorationand may lead to excessive deformations and crack widths, debonding ofinternal reinforcement in concrete members and rupture of the reinforce-ment and/or the concrete. For post-tensioned prestressed beams, fatiguefailure may occur at the anchorages. However, for a post-tensioned pre-stressed concrete beam for which precompression is such that the section

SOR7 12/19/98, 3:40 PM200


remains uncracked under service load cycling, fatigue does not generallycause significant problems. An important property of materials subjected todynamic fatigue loading is the ratio of the stress level at which irreversibledamage occurs to the stress for complete fracture of the material. The stressat which fatigue failure occurs after a given number of loading cycles isreferred to as the ‘fatigue strength’.

7.11 Fatigue of unplated beams

A typical reinforced concrete bridge deck may experience up to 7 � 108

stress cycles during the course of its 120 year lifespan, thus it is important tobe able to assess the fatigue performance of such structures. Concrete, steeland reinforced concrete members are all susceptible to fatigue loading.During the fatigue of concrete, irreversible deformation in the form ofcracking occurs; this is extensive and causes higher strains at failure thanwould occur under a static load. Cyclic loading at low levels (below thefatigue limit) improves the subsequent fatigue performance (when loadedabove the fatigue limit) of concrete and also the static strength (an increaseof 5–15%). The initial low stress cycling is thought to cause a densificationof the concrete.

As concrete ages its fatigue strength increases at a similar rate to its staticstrength. Consequently, for a given number of cycles, failure due to fatigueoccurs at the same percentage of ultimate strength. Failure at the bondbetween the cement paste and the aggregate is thought to be the predomi-nant failure mode in fatigue. Fatigue specimens tend to exhibit less brokenaggregate particles than comparable static test specimens.

However, it is the behaviour of the reinforcing steel that tends todominate the behaviour of conventionally reinforced concrete structures.The Transport and Road Research Laboratory (TRRL) undertook a pro-gramme of research (Moss, 1982) into the fatigue behaviour of the rebars,both in air and within reinforced concrete beams.

Axial tests of rebars can conveniently be undertaken utilising small cou-pon specimens in tension; the test frequency used can be high and thus datacan be rapidly obtained. However, the test conditions are not representa-tive of those rebars in reinforced concrete, consequently, tests on reinforcedconcrete beams are necessary. With these beams the testing frequencyneeds to be kept relatively low (in the TRRL case 3 Hz) to avoid hysteresiseffects. Such tests are time consuming and expensive.

TRRL used concrete beams, 3400mm long, 120mm wide and 220 mmdeep, spanning 3000 mm and loaded in four point bending at the thirdpoints. The beams utilised C55 concrete and were reinforced in the tensilezone with a single T16 bar.

In the analysis of the results the following relationship was derived:

SOR7 12/19/98, 3:40 PM201


Nσrm � K

where σr � stress range, N � cycles to failure, m � inverse slope of logσr �logN curve � 8.7, K�K0 gives the mean line of the relationship � 0.11 �1029 or K�K2 gives the mean minus 2 standard deviation line � 0.59 � 1027.

Fatigue failure of the reinforcement was found to occur at the position ofa crack in the concrete and the crack initiation site in the steel was associ-ated with the rib pattern on the surface of the bar.

The fatigue of a reinforcing bar in direct tension was found to be differentfrom that of a bar in the tension zone of a reinforced concrete beam. A barembedded in a concrete beam and loaded in flexure is more highly strainedin the section furthest from the neutral axis of the beam and, in addition, asthe concrete cracks in the tension zone of the beam, high strain values andhence high stresses in the bar spanning the crack will be developed. How-ever, as the maximum stress occurs at only one location there is less chanceof this coinciding with a defect in the bar. This occurrence gives an en-hanced fatigue life to a rebar embedded in a concrete beam compared witha bar cycled in air in direct tension.

As with other engineering properties the geometric scale is an importantconsideration in testing. Experimental data generated from laboratory testson small steel specimens differ from those derived from ‘life size’ specimensor the simulated testing of full scale structures. The larger the specimen, thelower the fatigue limit. Thus small laboratory specimens can give artificiallyhigh values of safe working stress levels.

Earlier work at RMCS (Emberson and Mays, 1996) examined the fatiguebehaviour of repaired concrete beams. Twelve beams 2540mm long by230mm deep by 150 mm wide were repaired either on the tension orcompression face. Each beam was subjected to a cyclic load, between 3 and30kN at a frequency of l Hz, in four point bending. Endurance of therepaired beams ranged from 70–280% of the control beams (average of590000 cycles to failure). In each case the failure was caused by tensilefatigue fracture of the main reinforcement. This corresponds to the overallconclusion that the fatigue behaviour of reinforced concrete beams isgoverned by the steel reinforcement.

7.11.1 Environmental effect on fatigue behaviour

The fatigue behaviour of reinforced concrete under environmental liveloading from wind and waves was an important aspect in a research pro-gramme called ‘Concrete in the Oceans (CiO)’ undertaken in 1989 for theDepartment of Energy (Leeming, 1989). A comprehensive review of exist-ing fatigue data was undertaken and a computer database of publishedresults was compiled. Of specific interest to the programme were crack

SOR7 12/19/98, 3:40 PM202


blocking and corrosion. Crack blocking occurs in concrete beams testedcyclically in seawater where deposits build up in the cracks giving rise to areduction in the deflection range and hence an increased fatigue life. Corro-sion can cause both initiation of the fatigue crack and blunting of the fatiguecrack by intense local corrosion. The effect of reinforcing bar size was alsoinvestigated; however, there was no clear trend from tests on 16mm, 32 mmand 40 mm bars in concrete beams. Following a programme of fatigue testson reinforced concrete beams a design curve was derived.

7.12 Fatigue of adhesives

Much of the research into the fatigue performance of adhesive bonds hasbeen carried out with aerospace applications in mind. Pioneering work wascarried out in Germany in the 1960s by Matting and Draugelates (1968)when a vast amount of data was collected using single lap joints of sheetmetal alloys bonded with phenolic and epoxy adhesives. The results werepresented in the same way as used for metal fatigue, that is in the form ofS–N curves of stress against number of cycles to failure using logarithmicscales. With metals it is generally accepted that endurance is dependent onthe stress range applied during cyclic loading, although with epoxy adhe-sives there is some evidence to suggest that crack growth rate may bedependent on the maximum stress intensity factor in the cycle.

This early work also provided evidence that the S–N curve is not verysensitive to frequency. However, as adhesives become more ‘rubbery’, forexample as they pass through the transition temperature, Tg, the time tofailure becomes more important thus implying different failure mecha-nisms; the mechanism before the Tg is reached is associated with crackgrowth, whilst the failure mechanism, for situations with values greater thanthe Tg, is a rate process. For both phenolic and epoxy adhesives, the endur-ance limit, defined as the maximum stress at which the survival time isindefinite, was found to be at about 15% of the static strength.

Other work (Marceau et al., 1978) using aluminium/epoxy lap joints at arange of frequencies showed that lower frequencies are more severe interms of the mean damage sustained per cycle. The difference in fracturemodes observed at different frequencies supports the theory that two differ-ent failure mechanisms are operating, that is a fatigue separation at highfrequencies compared with a creep–rupture separation at low frequencies.Elevated temperatures were also observed to shorten fatigue life but theeffect of humid air was less significant.

In the UK, Wake et al. (1979) attempted to establish the reality of anendurance limit based on the argument that successive cycles of stressabove the limit caused the growth of cracks from existing flaws. Their worksuggests an endurance limit close to a peak loading corresponding to 35%

SOR7 12/19/98, 3:40 PM203


of static ultimate strength. A small study was also conducted to assessthe relative importance of the crack initiation and propagation stages onsubsequent strength. It was concluded that, if an endurance limit exists,load cycling below it may only serve to initiate cracks (not detected bysubsequent static tests) rather than to propagate them. Failure may then beinitiated should cycling above the endurance limit subsequently take place.

Much of this early research and its subsequent development relates totypes of adhesive and substrates unlikely to be encountered in civil engi-neering applications. For this reason independent work was commissionedby the TRRL, now TRL (Mays, 1990). The test programme used steel-to-steel double lap joints manufactured with one of two cold cure epoxiestypical of those used in structural engineering. These were subject to cyclicloading at a frequency of 25Hz at load ranges to produce failure atendurances of up to 100 million cycles. Fatigue tests were undertaken onspecimens cured and weathered under a range of climatic conditions andtested at temperatures ranging from �25 °C to �55 °C.

The results are presented in Fig. 7.9 in terms of stress range rather thanpeak stress, to be consistent with accepted engineering practice for metals.The correlation obtained using this approach was quite acceptable, al-though the quoted values of stress range will be somewhat sensitive togeometrical configuration of the joint employed. Although the results lendcredence to the existence of an endurance limit, its value has not beenprecisely determined.

Compared with S–N curves for metal fatigue, those for adhesive bondedjoints are relatively flat, that is, there is a large improvement in endurancefor a relatively small reduction in stress range. From an engineering point ofview, this is advantageous since, in the absence of an endurance limit, the S–N curve may be extrapolated to determine the stress range necessary toestablish the maximum conceivable number of repetitions of a constantlyapplied load.

For the cold cure adhesives utilised in this study, the lower bound tofatigue performance for service temperatures between �25 °C and �45 °Cis represented by the relationship:

NSr � K

where N � cycles to failure, Sr � stress range, K � 2 � 1022 for mean minus2 standard deviation line.

Extrapolating to 7 � 108 cycles, representing the maximum conceivablenumber of significant load variations during the 120 year design life of abridge, yields a limiting stress range of 4.0Nmm�2. This figure approximatesto the tensile shear strength of the concrete in applications involving rein-forced concrete members with externally bonded reinforcement. However,typical adhesive shear design stresses at the serviceability state are unlikely

SOR7 12/19/98, 3:40 PM204


205

Figure 7.9 Lower bound to fatigue performance of cold cure epoxies (Mays, 1990).

SO

R7

12/19/98, 3:40 PM

205


to exceed 1.0–1.5Nmm�2 so that fatigue is unlikely to be a critical designconsideration at these temperatures.

The brittle failure modes of joints made with these cold cure adhesivesfor service temperatures up to 45°C confirm the prediction that a mecha-nism dependent on crack growth is predominant. However, at 55°C, atemperature above the Tg, the more ductile failure mode suggests a signifi-cant change in material response which is reflected in a dramatic reductionin fatigue performance. Thus, the use of cold cure adhesives such as these isnot recommended in situations where the service temperature exceeds45 °C or Tg �10 °C, whichever is the lower.

7.13 Fatigue of FRP materials

One of the significant advantages of composite materials over metals istheir superior fatigue performance. Composites have traditionally foundapplications in the aerospace industry, in which their uses have reflectedtheir superior fatigue resistance. Steel usually fails by the sudden propaga-tion of a single crack at the end of the fatigue life, but polymeric compositesexperience progressive fatigue degradation due to failure of the fibres andthe matrix and of the matrix/fibre interfaces (Ellyin and Kujawski, 1992).

Unlike isotropic materials, the cyclic degradation of anisotropic materi-als, in the form of microcrack growth, delamination and subsequent mixedmode crack propagation, may be well established before complete fractureoccurs. A considerable degree of experimental scatter is found in the litera-ture, due partly to uncertainty about the moment in time at which a com-posite can be considered to have failed (Dharan, 1975). The progressivedegradation of composites renders their change in modulus a more impor-tant design criterion than for a homogeneous material, but the fatigueperformance of composites is likely to be more than adequate due to therelatively slight static constraints. The scatter in static strengths of poly-meric composite materials leads to low design stresses being adopted inpractice; these low working stresses may result in the elimination of fatigueas a design criterion.

The slopes of the S–N curves (i.e. the measure of relative fatigue per-formance) of composites are determined principally by the strain in thematrix, since the fatigue limit of the matrix is lower than that of the fibres(Curtis, 1989). The fatigue behaviour of composites containing variousresins is not very different despite great chemical differences, although thefatigue properties of epoxies are slightly superior due to their greaterstrength, higher resin/fibre interface strengths, lower curing shrinkagewhich implies lower residual stresses before loading, and high strain tofailure which prevents fibre exposure and delays stress corrosion in GFRPs(Dew-Hughes and Way, 1973).

SOR7 12/19/98, 3:40 PM206


The four basic fatigue failure mechanisms of polymeric composites arematrix cracking, delamination, fibre fracture and interface debonding(Hollaway, 1993). The type and degree of these mechanisms vary depend-ing on material property, laminate stacking sequence and the type of fa-tigue. Matrix cracking in the off-axis plies of multidirectional laminates isusually the first damage mechanism because the lack of fibres oriented inthe direction of the applied load causes more load to be distributed to thematrix. The slope of the tensile S–N curve increases with reducing staticstrength as the proportion of off-axis fibres increases; this arises due tothe increasing likelihood of matrix damage with a reducing proportionof fibres oriented in the direction of the applied load. The final fracture ofmultidirectional composites occurs when the 0° fibres fail. The fibres ofunidirectional composites carry the majority of the applied load and thesecomposites have excellent fatigue resistance as a result (Curtis, 1989).Delamination is attributed to the presence of interlaminar stresses inthe vicinity of a free edge under in-plane loading (Hollaway, 1993) givingrise to damage propagation from the free edge inwards. Renton andVinson (1975) found that the laminate stacking sequence will affect thefatigue failure mode of bonded composite adherends; unidirectional com-posites caused cohesive failure in the adhesive, while an alternating �45°sequence, with a 45° layer against the adhesive, resulted in compositedelamination.

The critical stage in the fatigue failure of composites is the initial resin/fibre debonding. Dew-Hughes and Way (1973) cited results which showedthat fibre/matrix debonding in GFRPs may start at 30% of the static tensilestrength and at decreasing stresses under cyclic loading, while failures dueto the cyclic loading of CFRPs may be delayed until 70% of the staticstrength. The existence of debonding at a low proportion of the staticstrength implies damage in the first cycle under dynamic fatigue condi-tions, as confirmed experimentally by Kim and Ebert (1978) for GFRPcomposites.

Statistically distributed flaws in the fibres of a unidirectional CFRP speci-men may cause fibres to break before resin damage appears; a crack willextend into the resin matrix with further increasing load and its path willdepend on the strength of the fibre/matrix bond of the fibres towards whichthe crack travels. A strong fibre/matrix bond causes cracks to extend intothe matrix, while low interface strengths result in interfacial debond andfibre pull-out (Hollaway, 1993). Also, the elastic modulus of carbon fibresmay be as much as six times greater than the modulus of glass fibres, so thefracture strains of carbon fibres are relatively low, certainly below thestrains at which common resin matrix materials begin to craze (Beaumontand Harris, 1971). The fracture strains of glass fibres, however, are greaterthan the strains to first matrix damage.

SOR7 12/19/98, 3:40 PM207


The rate of matrix and interfacial damage propagation is affected by thebulk strain in the resin matrix. The matrix strain is, in turn, dependent onthe modulus of elasticity and the volume fraction of the fibres. High modu-lus carbon fibres result in relatively low strains and, therefore, lower rates ofdamage propagation than in composites containing lower modulus glassfibres (Curtis, 1989). In regions of a composite with closely spaced fibres(i.e. if the fibre distribution throughout the cross-section is not uniform),resin cracks may propagate by the coalescence of individual fibre/matrixdebonds (Hull, 1992). When the fibres are widely spaced, the growth of acrack between fibres depends on the fatigue crack growth resistance of thematrix.

Little improvement in fatigue behaviour is generated by the use of rein-forcing fibres, of a particular type, with a higher static strength in a givenresin matrix, since the fatigue failure is usually governed by damage in thematrix and at the fibre/matrix interfaces so the improved fibre strain will notbe used to advantage (Owen, 1974).

The effect of environmental exposure on the fatigue behaviour of com-posites depends on the sensitivity of the laminate to the properties of thematrix with respect to environmental conditions (Curtis, 1989). This isbecause it is usually the matrix or fibre/matrix interface that is affected bymoisture absorption, as noted in Chapter 6. A high proportion of the totalmoisture uptake reaches the fibre/matrix interfaces, as seen in the smallweight gains of unreinforced resins exposed to steam for 200h, comparedwith the weight gains of CFRPs under the same exposure (Beaumont andHarris, 1971). GFRPs are affected by an increase in moisture content due inthe main to the stress corrosion of the fibres, so the influence is smaller incomposites under lower stress (Dew-Hughes and Way, 1973).

The high fibre/matrix interface bond strengths of CFRPs produce littleoverall fatigue sensitivity to moisture at room temperature (Owen, 1974).As for static strengths, the fatigue strengths of composites are degradedmost by the combined influence of elevated temperature and relative hu-midity. In addition to providing relatively high static and fatigue strengths,a unidirectional fibre orientation results in the best environmental durabil-ity under cyclic loading, since the fibre/matrix interfaces are not loaded intension normal to the fibres (Beaumont and Harris, 1971; Demers, 1997a;Demers, 1997b).

7.14 Fatigue of plated beams using steel plates

The early EMPA work (Ladner and Weder, 1981) included fatigue tests onplated reinforced concrete beams. Initial tests on small rectangular cross-section beams (150 � 250 � 2400 mm) were followed by tests on a ‘T’ beam(500 � 900 � 6700mm).

SOR7 12/19/98, 3:40 PM208


For the rectangular beams, a small static load was applied in orderto initiate cracking prior to plating. A frequency of 4.3Hz was selected,with a maximum of 107 load cycles, the load levels being varied for eachof the eight beams tested. Steel stress ranges in the plate at midspanvaried from 120 � 20 Nmm�2 to 150 � 130Nmm�2. Three beams survived107 cycles, in two beams fracture of the steel plate occurred and the re-maining three beams failed due to fatigue of the bond. In these casesthe steel plate debonded from the adhesive at crack locations in theconcrete and this debond crack then propagated towards the end of theplate.

The ‘T’ beam was strengthened with both a flexural plate on the soffit andshear plates on the side of the beam. The beams were subjected to fourphases of 2 � 106 cycles, each subsequent phase consisting of an increasedload range. One of the internal rebars failed near the end of the secondfatigue phase. Testing continued with the span reduced from 6000 mm to4900 mm. Following cyclic loading, the beam was tested statically and thisshowed that the beam had survived the fatigue phases without any damage(excepting the rebar fracture mentioned above).

In 1985, under-reinforced concrete beams (100mm � 100 mm �

1000 mm) were strengthened using three different adhesives to bond onsteel plates and subsequently tested in fatigue (Mays, 1985). Testing wascarried out for up to 2 � 107 cycles at a frequency of 6Hz. There was nodiscernible difference in performance for the three different adhesives.

Fatigue failures occurred by progressive horizontal cracking within theconcrete at the level of the internal reinforcement and eventual platedebonding at the concrete/adhesive interface. The failure mode was domi-nated by the stress concentration occurring in the adhesive close to the plateends. Within the adjacent concrete, compatible stresses led to cracking andhence a shift in the position of the stress concentration. This progressiveconcrete cracking and plate debonding eventually led to failure. Beams thatwere precracked prior to plating failed due to fracture of the soffit plateadjacent to one of the flexural cracks.

When compared with an unplated control specimen, an improvement infatigue performance was evident. At a load range of 18.9kN the controlbeam failed at 257 000 cycles compared with 6.6 � 106 cycles for a platedbeam at the same load range.

Overall, no relevant fatigue damage occurred with the tensile stress(from the measured plate strain) in the soffit plate below 150N mm�2. It wasconcluded that as the peak stress in the external plate reinforcement rarelyexceeds 100 Nmm�2 under typical design loading then fatigue is unlikely tocause concern.

Further fatigue testing of strengthened reinforced concrete beams wasundertaken at the EMPA Laboratories in Switzerland (Holtgreve, 1986).

SOR7 12/19/98, 3:40 PM209


Reinforced concrete ‘T’ beams, 600mm deep and spanning 7000 mm, wereinitially loaded to induce cracking. Steel plates were then bonded to thesoffits of the beams whilst they were under dead load. Following curing ofthe adhesive, the beams were cycled at a rate of 4Hz from 117–254 kN.After 2 � 106 cycles the cyclic loading was stopped. There was no evidenceof any bond deterioration between the steel and the concrete and the beamswere then subjected to a static ultimate load test. Typical failure loads of436kN were achieved, with failure occurring in the outer tensile zone of theconcrete, no failure of the adhesive being evident. Maximum strains of0.2% were recorded in the steel plate.

Welded plates are not recommended for use in steel plate bondingand their susceptibility to fatigue loading has been demonstrated(Canovas, 1990). Rectangular section reinforced concrete beams 200mmdeep, 250 mm wide and spanning 3000mm were strengthened withadhesively bonded steel plates some of which were joined by welding. Thebeams were tested at 8.3Hz for 2 � 106 cycles between 50% and 75% ofthe ultimate bending moment and a further 2 � 106 cycles between 60%and 80%. The results showed that some adhesives are more susceptible tofatigue than others and that welded steel plates are very susceptible tofatigue.

The mode of failure arising from dynamic loading has been carefullystudied (Hankers, 1990). Dynamic tests were carried out on plated rectan-gular reinforced concrete beams, with a cross-section of 260mm � 150 mm,spanning 1800 mm. Measurements of load, deflection, plate strains, rebarstrains and visual inspection were performed as the beams were tested at afrequency of 4Hz.

At the high load levels, failure tended to be by debonding. Starting frombending cracks near the central load application point, the crack (betweenthe steel plate and the concrete) proceeded towards the ends of theunanchored plate and this led to sudden failure. At the lower load levels theunfailed beams were tested statically, producing higher load capacities thanthe statically tested control beams. Possible vibration-activated adhesivepolymerisation was suggested.

Rectangular cross-section beams (300 mm � 200mm � 3000mm)were again used in the most recently reported study on the dynamic behav-iour of reinforced concrete beams strengthened with externally bondedsteel plates (Taljsten, 1994). Two beams were tested dynamically in fourpoint bending, with a load of 75 � 55 kN and at a frequency of 3 Hz. Theresults of this and the previous research suggest that plated beams canbehave better than unplated beams when subjected to cyclic loading,possibly because the bonded plates delay the initial cracking and reducecrack widths.

SOR7 12/19/98, 3:40 PM210


7.15 Fatigue of short span plated beams using

FRP plates

As with the long term sustained load testing of plated beams, the fatiguebehaviour of plated concrete members has received much less attentionthan the testing of more convenient small scale lap shear specimens.

Fatigue tests of beams strengthened by bonded CFRP plates wereundertaken at EMPA (Kaiser, 1989; Deuring, 1993) and described inEnglish in various publications (Meier and Kaiser, 1990; Meier and Kaiser,1991; Meier et al., 1993a; Meier et al., 1993b). For the T-beams at a loadingfrequency of 4Hz, the first fatigue failure occurred in one of the twointernal bars at 4.8 � 105 cycles, the second bar broke at 5.6 � 105 cycles,the first bar broke at another location after 6.1 � 105 cycles and thesecond bar broke again at 7.2 � 105 cycles, compared with the first externaldamage of the CFRP plate at 7.5 � 105 cycles and complete failure ofthe plate at 8.05 � 105 cycles, indicating that the plate was able tosustain significant further fatigue loading after failure of the internalreinforcement.

Fatigue tests by Shijie and Ruixian (1993) showed that the fatigue lives ofGFRP plated beams were three times longer than the lives of comparableunstrengthened beams. The fatigue strengths were increased by between15% and 30% and the midspan deflections reduced by 40%; the maximumcrack widths and crack propagation were also reduced. The postcyclic staticstrength and stiffness diminished with increasing numbers of cycles but bya smaller magnitude than in unstrengthened beams.

Chajes et al. (1995a) used Sikadur 32 adhesive to bond aramid, carbonand glass fibre composite plates to small scale reinforced concrete beams inwhich the plate had the same tensile strength as the single internal rebar,achieved by varying the number of plate layers of the three types. Thebeams, which were still under test at the time of writing, were to be sub-jected to 70% of their static ultimate capacity for 2.5 � 106 cycles and thentheir postcyclic static capacities were to be compared with the results of thebeams tested by Chajes et al. (1995b).

In the ROBUST programme, five 1.0 m long beams were plated with aprepreg CFRP composite material, before being load tested cyclically at afrequency of 1 Hz, between applied loads which generated tensile stresses inthe internal mild steel round rebars ranging between 2% and 90% (at thelower and upper loads, respectively) of the yield stress. The correspondingvalues for the unplated beams were 3% and 59%. All the beams wereloaded to an upper level of approximately half of their ultimate capacity,representing 132% of the serviceability load in the plated cases. The tensilestress range in the rebars was important since reinforced concrete beams

SOR7 12/19/98, 3:40 PM211


have typically been found to fail due to the fatigue fracture of the reinforc-ing steel under cyclic loading (Canovas, 1990; Bannister, 1969; Lovegrove,1979; Tilly, 1979; Tilly and Moss, 1982). This was the case in the unplated1.0m beam tested in the ROBUST project in which two out of three rebarsfailed after about 1.5 � 106 cycles. Figure 7.10 shows the deflection ampli-tude against the number of cycles of load for one of the plated beams.Although the stress in the rebars was greater in plated beam compared withthe unplated one, no rebar damage was recorded up to 11 � 106 cycles atwhich point the test was stopped. This result reflected the ability of the plateto limit the concrete crack widths.

In the fatigue tests undertaken on the 1m long beams in the ROBUSTproject, it was shown conclusively that, under a shear span/beam depthratio of between 3.4 and 4.0, it is necessary to incorporate a bolted plateend anchorage system. It was explained in Chapter 4 that under a staticload it is necessary to install an anchorage system at the free end of theplates to gain the maximum strength and stiffness of the plated beamsystem. Furthermore, it was also concluded that the bolts in the anchorsystem should be located nearer to the free end of the anchorage block.This will increase the bond length over which the plate tension is trans-ferred to the bolts and the plate splitting problem, which occurs when thebolts are positioned in the centre of the plate, will be eliminated. In allanchorage systems, a suitable adhesive for bonding the anchorage block tothe composite must be used; in the ROBUST project the 3M adhesive wasemployed.

Figure 7.10 Deflection amplitude variations of unplated and platedbeams (Garden, 1997).

SOR7 12/19/98, 3:40 PM212


The fatigue performance of the composite material is usually excellentso this will not be a design issue. Even under an upper cyclic plate stressrepresenting 50% of the plate strength, using a prestressed composite inthe ROBUST work, the CFRP plate material exhibited no indication ofdamage.

7.16 Fatigue of long span plated 2.3m beams using

FRP plates

The ROBUST project continued by undertaking research into thefatigue, at the RMCS, of long span plated beams using FRP plates.The aim of the experimental programme was to provide initial designguidance on the fatigue performance of reinforced concrete beamsstrengthened with CFRP plates. Six reinforced concrete beams, 2300mmlong, 130mm wide and 230 mm deep were used for this study, similar sizedbeams being utilised in the static testing regime discussed in Chapter 4.Reinforcement consisted of three number T12 bars in the tension zoneat an effective depth of 205mm, two number T8 top bars and R6 links at150 mm centres to ensure a flexural failure mode. All the steel was tiedrather than welded in order to avoid fatigue failures of any welds. Thebeams were cast using grade C50 concrete. Four of the six beams werestrengthened by bonding the CFRP plate developed in the ROBUSTproject.

For testing, the beams were supported on a purpose built rig whichpermitted the beams to behave in a simply supported manner but whichalso limited the amount of longitudinal and transverse movementwhich could develop during the test. Load was applied via a servohydraulicactuator to a spreader beam and hence to the beam itself through rollers inorder to achieve four point bending. The beams spanned 2100 mm and wereloaded at two points, each 130 mm from the centre span. Figure 7.11 showsthe testing arrangement.

Prior to the fatigue loading, a static load test was undertaken up to themaximum load to be used in the subsequent fatigue loading cycle. Eachbeam was subjected to a cyclic load in four point bending at a frequency of1 Hz. This frequency was selected on the basis of previous research(Emberson and Mays, 1996).

For the static test, the midspan deflection and the strains across thesection at midspan were measured, using a linear variable differential trans-former (LVDT) and a demec gauge, respectively, at each load increment.During the fatigue test, the load, midspan deflection and number of cycleswere monitored. At 200 min intervals, 25 data points were recorded, at arate of 20 data points per minute. For beam number six, the strain in theconcrete and the CFRP plate was also recorded.

SOR7 12/19/98, 3:40 PM213


Figure 7.11 Long span beam fatigue testing apparatus.

7.16.1 Results of the 2.3m long beams tested underROBUST

Six beams, two unplated control beams and four plated beams, were tested.The load conditions are detailed in Table 7.2. The main objective of theprogramme was to obtain data that will enable an understanding to begained of the fatigue performance of reinforced concrete beams strength-ened with CFRP plates. This incorporates the dual aims of:

1 representing the fatigue loading on reinforced concrete structures,2 comparing the fatigue life and behaviour of an unplated beam and a

comparable plated beam.

The beam types to be discussed below will be referred to as beams 1 to 6,where beams 1 and 2 are the unplated beams and beams 3, 4, 5, and 6 are theplated members. When a fatigue load investigation is undertaken there arevarious loading options which can be applied to the beams being examined;the options chosen in the ROBUST project are explained below.

1 To apply the same loads to both the unplated and plated beams. Inmany cases in practice, a degraded beam needs to be strengthened to itsoriginal capacity. Beams 1 and 4 were cycled under the same loadingmagnitudes.

2 To apply loads to both types of beam to give the same stress in theirrespective rebars. In order to make a comparison, the unplated andplated beams should be tested at the same percentage of ultimate stress

SOR7 12/19/98, 3:41 PM214


Table 7.2 Loading conditions and results summary of fatigue tests at RMCS on2.3m beams

Test Type of Predicted Load Load No. of Max. Stress Failureno. beam ultimate range range1 cycles stress2 range (%) mode

capacity (kN) (kN) (MPa)

1 Unplated 75 4–40 48 20 000 54.8 49.3 Steelyield

2 Unplated 75 3–32 39 732 600 43.7 39.3 Steelyield

3 Plated 113 5–49 39 508 500 54.8 49.3 Steelfracture

4 Plated 113 4–40 32 321 195 43.7 39.3 Steelfracture

5 Plated 113 4–40 32 1 889 087 43.7 39.3 Steelfracture

6 Plated 113 2.9–32 25.9 12 � 106 36.3 33.0 Teststopped

1 % of ultimate capacity.2 % of ultimate stress in steel bar at maximum load.

in the steel at maximum load and not at the same stress value. Beams 1,2 and 3, 5 were tested with comparable values of percentage of ultimatestress in the steel rebars at maximum load.

3 To apply an equal percentage of the ultimate load capacity of eachrespective beam. If an unplated beam were tested at 50% of its ultimatecapacity, the plated beam should be tested also at 50% of its ulti-mate capacity. Beams 2 and 3 were tested at comparable values ofultimate capacity.

Beam 4 failed after a surprisingly low number of cycles and the loadconditions were repeated for beam 5. An examination of the static testresults showed that beam 4 was considerably less stiff than the other beams,suggesting a possible weakness present in this beam which may have led topremature failure. In general, however, the static test results showed thatthe unplated beams (beams 1 and 2) are less stiff than the plated beams(beams 3, 5 and 6), as would be expected.

Typical fatigue duration on reinforced concrete structures is in the regionof 107 cycles or more. However, within the confines of the programme, itwas not possible to conduct cycling on each beam for such a long period oftime (107 cycles at 1Hz is about 4 months).

The results are summarised in Table 7.2 and are also displayed graphi-cally on the S–N curves displayed in Fig. 7.12 alongside the curves forreinforced concrete beams obtained by TRRL (Moss, 1982) and Concretein the Oceans (CiO) (Leeming, 1989).

SOR7 12/19/98, 3:41 PM215


7.16.2 Discussions of the 2.3m span beams usingFRP plates

The fatigue results will now be assessed by referring to the three loadoptions described above. Considering each of these in turn:

1 Beams 1, 2, 5 and 6 were cycled under the same load. The strengthenedbeam, in each case, had a considerably enhanced fatigue life.

2 Beams 1, 2, 3 and 5 were tested with comparable values of percentage ofultimate stress in the steel bar at maximum load. Again the platedbeams exhibit an enhanced fatigue life.

3 Beams 2 and 3 were tested at the same percentage of ultimate capacity.The plated beam had a shorter fatigue life than the unplated beam.

From the above assessment of the results it would be unwise to expect thesame fatigue capacity, in terms of percentage of ultimate capacity, from aplated beam as from an unplated one. A more reasonable value for designguidance would be to expect the same fatigue life for plated and unplatedbeams with comparable values of percentage of ultimate stress in the steelbar at maximum load.

A study of the failure mechanisms was undertaken, the results of whichare described below:

• Beam 1: there was no evidence of failed rebars. It is assumed that failurewas by yielding of reinforcement.

• Beam 2: the load dropped to a value of 20kN after 732 600 cycles. At744 660 cycles, the beam was unable to sustain an increase in load to

Figure 7.12 S–N curve for fatigue of CFRP plated beams comparedwith TRRL (Moss, 1982) and CiO (Leeming, 1989) curves.

SOR7 12/19/98, 3:41 PM216


enable the original value of 32kN to be restored. It appeared that thesteel had yielded but no sign of rebar damage was found.

• Beam 3: at 508 500 cycles, a large deflection occurred, presumably due tofracturing of the main internal reinforcement. The cyclic load continuedat a low value (0–20kN) with only a partially bonded CFRP plate. Thecyclic load stopped after 584100 cycles, when the bottom steel fractured.

• Beam 4: at 321 200 cycles, a large deflection occurred, presumably dueto fracturing of the main reinforcement. The cyclic load continued at alow value (0–20kN) with only a partially bonded CFRP plate. When anattempt was made to increase the load to its original value, the CFRPplate failed at the anchorage and the beam hinged about the top steel at568 200 cycles; the bottom steel fractured at this point.

• Beam 5: at 1889087 cycles, a large deflection occurred, presumably dueto fracturing of the main internal reinforcement. The cyclic loadingcontinued at low load with the main reinforcing steel fractured and theload being taken in the tension zone of the beam by the composite plate.When an attempt was made to increase the applied load to its originalvalue, the CFRP debonded, the deflection increased, the plate split andfinally pulled away from the end anchors. The beam hinged about thetop reinforcement and upon inspection it could be seen that the bottomsteel had fractured.

• Beam 6: during the initial static loading, the strain was highest atmidspan, dropping considerably towards the end of the plates. Duringthe fatigue test, an overall increase in strain at each location indicatedstress transfer to the plate as the beam cracked. There was a correspond-ing increase in compressive strain on the top surface of the concretebeam. There did not appear to be any stress redistribution with timealong the length of the plate. The test was concluded at a value of2.5 � 106 cycles.

7.17 Concluding summary

The component parts of an RC plated beam have all been shown to exhibitvarying degrees of creep deformation whilst under a constant load. Creeptests of plated beams strengthened by bonding CFRP composite plates onto their soffits have demonstrated that plated and unplated beams experi-ence similar proportional creep, suggesting that the creep behaviour of suchbeams is governed primarily by the compressive creep of the concrete. Theeffect of temperature upon the creep performance of adhesives dictates thatthe use of plate bonding at elevated temperatures should be viewed withcaution.

Composite materials exhibit superior fatigue performance to that of steeland from the research to date it would appear that it is the fatigue of the

SOR7 12/19/98, 3:41 PM217


reinforcing steel within a plated reinforced concrete beam that is the domi-nant factor in fatigue performance. Thus, it is recommended that for satis-factory fatigue performance the design of CFRP plated beams should limitthe performance of the ultimate stress in the steel reinforcement at maxi-mum load to that allowed in an unplated beam.

7.18 References

Aboudi J (1990) ‘Micromechanical characterisation of the nonlinear viscoelasticbehaviour of resin matrix composites’, Composites Sci Technol 38 371–386.

ACI (1982) Prediction of Creep, Shrinkage and Temperature Effects in ConcreteStructures, ACI Committee 209, ACI Detroit.

Adams R D and Wake W C (1984) Structural Adhesive Joints in EngineeringLondon, Elsevier Applied Science.

Bannister J L (1969) ‘The behaviour of reinforcing bars under fluctuating stress’,Concrete 3, 405–409.

Barnes R A and Mays G C (1988) Creep of Adhesive Joints at Elevated Tempera-tures, RMCS Contractor’s Report to TRRL, July 1988.

Beaumont P W R and Harris B (1971) ‘The effect of environment on fatigue andcrack propagation in carbon fibre reinforced epoxy resin’, in Carbon Fibres, TheirComposites and Applications, Proceedings International Conference, organisedby the Plastics Institute, London, Feb 2–4, 1971, pp 283–291.

Calder A J J (1990) ‘The durability of steel plates bonded to concrete with structuralepoxy adhesives’, International Seminar Structural Repairs/Strengthening by thePlate Bonding Technique, University of Sheffield, September 1990.

Canovas F M (1990) ‘Strengthening of structural elements of reinforced concrete bymeans of glued steel plates – some applications in Spain’, International SeminarStructural Repairs/Strengthening by the Plate Bonding Technique, University ofSheffield, September 1990.

Chajes M J, Thomson T A and Farschman C A (1995a) ‘Durability of concretebeams externally reinforced with composite fabrics’, Construct Building Mater9(3) 141–148.

Chajes M J, Thomson T A and Tarantino B (1995b) ‘Reinforcement of concretestructures using externally bonded composite materials’, in Non-Metallic (FRP)Reinforcement for Concrete Structures, ed. L. Taerwe, London, E & FN Spon.

Cottingham W S, Fluck P G and Washa G W (1961) ‘Creep of prestressed concretebeams’, ACI J 57 929–936.

Crossman F L and Flaggs D L (1978) Lockheed Missiles and Space Center, ReportNo D633086.

Curtis P T (1989) ‘The fatigue behaviour of fibrous composite materials’, J StrainAnal 24(4) 235–244.

Demers C (1997a) ‘Tension–tension axial fatigue of carbon fiber reinforced poly-meric composites’, in Recent Advances in Bridge Engineering. US – Canada-Europe Workshop on Bridge Engineering, Zurich, July 1997, pp 170–177.

Demers C (1997b) ‘Tension–tension axial fatigue of E-glass fiber reinforced poly-meric composites, fatigue strength and fatigue models’, Structural Faults andRepair 97, Edinburgh July 1997, pp 175–182.

SOR7 12/19/98, 3:41 PM218


Deuring M (1993) ‘Verstärken von Stahlbeton mit gespannten Faserverbund-werskstoffen’, EMPA Bericht Nr. 224.

Dew-Hughes D and Way J L (1973) ‘Fatigue of fibre-reinforced plastics: a review’,Composites, Part A 4 167–173.

Dharan C K H (1975) ‘Fatigue failure mechanisms in a unidirectionally reinforcedcomposite material’, in Fatigue of Composite Materials, Proceedings of a sym-posium presented at December committee week of the American Society forTesting and Materials, Bal Harbour, 1975.

Dillard D A and Brinson H F (1983) ‘A numerical procedure for predicting creepand delayed failures in laminated composites’, in Long-Term Behaviour ofComposites, ed. T K O’Brien, ASTM Special Technical Publication No. 813,1983, Philadelphia, pp 23–37.

Ellyin F and Kujawski D (1992) ‘Fatigue testing and life prediction of fibreglass-reinforced composites’, in 1st Int. Conf. Advanced Composite Materials in Bridgesand Structures, eds K W Neale and P Labossière, Sherbrooke, Canada, 1992,pp 111–118.

Emberson N K and Mays G C (1996) ‘Significance of property mismatch in thepatch repair of structural concrete. Part 3: Reinforced concrete members inflexure’, Mag Concrete Res 48(174) 45–57.

Evans R H and Kong F K (1996) ‘Creep of concrete in R C and P C design’, CivilEng Public Works Rev 61(718) 594–596.

Ferry J D (1980) Viscoelastic Properties of Polymers, 3rd edn. John Wiley and Sons,New York.

Findley W N (1960) ‘Mechanism and mechanics of creep of plastics’, SPE J 57–65.Garden H N and Hollaway L C (1997) ‘An experimental study of the strengthening

of reinforced concrete beams using prestressed carbon composite plates’, 7th IntConf Structural Faults and Repair, Edinburgh, 1997.

Garden H N (1997) Part 2 Response of CFRP plated Reinforced Concrete beams tosustained load, Composite Materials, Internal Report, Department of Civil Engi-neering, University of Surrey, July 1997, pp 20–56.

Garden H N, Hollaway L C, Thorne A M and Parke G A R (1996) ‘A parameterstudy of the strengthening of reinforced concrete beams with bonded composites’,3rd Int Conf Bridge Management, Guildford, E & F N Spon, London, 1996.

Garden H N, Hollaway L C and Thorne A M (1997) ‘A preliminary evaluation ofcarbon fibre reinforced polymer plates for strengthening reinforced concretemembers’, Struct Buildings 122(2) 127–142.

Hankers C (1990) Strengthening of structural concrete members by bonded steelplates’, International Seminar Structural Repairs/Strengthening by the Plate Bond-ing Technique, University of Sheffield, September 1990.

Hiel C C, Brinson H F and Cardon A H (1983) ‘The nonlinear viscoelastic responseof resin matrix composites’, 2nd Int Conf Composite Structures, Paisley, Septem-ber 1983, pp 271–281.

Hollaway L C (1993) Polymer Composites for Civil and Structural Engineering,Glasgow, Blackie Academic and Professional.

Holtgreve K (1986) ‘A contribution in regard to behaviour of reinforced tee-beamswith bonded flatsteel components under fatigue loads, Proceedings InternationalConference on Adhesion between Polymers and Concrete, RILEM TechnicalCommittee 52, Aix-en-Provence, September 1986.

SOR7 12/19/98, 3:41 PM219


Howard C L and Hollaway L C (1987) ‘The characterisation of the nonlinearviscoelastic properties of a randomly orientated fibre/matrix composite’, Com-posites 18(4) 317–323.

Hull D (1992) An Introduction to Composite Materials, Cambridge, CambridgeUniversity Press.

Johnson A F (1979) Engineering Design Properties of GRP, Publication 215, BritishPlastics Federation, London.

Kaiser H (1989) Bewehren von Stahlbeton mit kohlenstoffaserverstärktenEpoxidharzen, Doctoral Thesis, Eidgenössischen Technischen HochschuleZürich, 1989.

Kim H C and Ebert L J (1978) ‘Axial fatigue failure sequence and mechanisms inunidirectional fiberglass composites’, J Compos Mater 12(April) 139–152.

Ladner M and Weder C (1981) Concrete Structures with External Bonded Reinforce-ment, EMPA Report 206, Dubendorf.

Lark R J and Mays G C (1984) ‘Epoxy adhesive formulation: its influence on civilengineering performance’, in Adhesion 9, ed. K W Allen, Elsevier AppliedScience, London, Chap 7, pp 95–111.

Leeming M B (1989) Concrete in the Oceans Programme – Coordinating Report onthe Whole Programme, London, HMSO.

Lovegrove J M, Salah El, Din A S and Daoud O K (1979) ‘Fatigue crack growth inthe tension steel of reinforced concrete’, in Fatigue Eng Mater 2(3) 173–183.

Marceau J A, McMillan J C and Scardino W M (1978) ‘Cyclic stress testing ofadhesive bonds’, Adhesives Age 21(4) 37–41.

Matting A and Draugelates H (1968) ‘Schwingfestigkeit von Metall-klebverbindungen (The vibrational stability of metal adhesive joints)’ Adasion11(1) 5–20, (3) 110–134, (4) 161–176.

Mays G C (1985) Fatigue Performance and Durability of Structural Adhesive Joints,PhD Thesis, University of Dundee.

Mays G C (1990) Fatigue and Creep Performance of Epoxy Resin Adhesive Joints,Transport and Road Research Laboratory Contractor Report 224.

Mays G C (1993) ‘The use of bonded external reinforcement in bridge strengthen-ing: structural requirements of the adhesive’, in Bridge Management 2, Proceed-ings of the 2nd International Conference on Bridge Management, eds J E Harding,G A R Parke and M J Ryall, University of Surrey, Guildford, April 18–21 1993,pp 672–680.

Mays G C and Hutchinson A R (1992) Adhesives in Civil Engineering, Cambridge,Cambridge University Press.

Meier U and Kaiser H (1990) ‘Strengthening of structures with CFRP laminates’,Proc. Int. Seminar Structural Repairs and Strengthening by the Plate BondingTechnique, University of Sheffield, 1990.

Meier U and Kaiser H (1991), ‘Strengthening of structures with CFRP laminates’,Proc. Specialty Conf. Advanced Composites Materials in Civil Engineering Struc-tures, Nevada.

Meier U, Deuring M, Meier H and Schwegler G (1993a) ‘CFRP bonded sheets’, inFiber-Reinforced-Plastic Reinforcement for Concrete Structures, Proc. Int. Sympo-sium, eds A Nanni and C W Dolan, Vancouver, 1993, ACI Special PublicationNo. 138.

Meier U, Deuring M, Meier H and Schwegler G (1993b) ‘Strengthening of struc-tures with advanced composites’, in Alternative Materials for the Reinforcement

SOR7 12/19/98, 3:41 PM220


and Prestressing of Concrete, ed. J L Clarke, Glasgow, Blackie Academic andProfessional.

Moss D S (1982) Bending Fatigue of High Yield Reinforcing Bars in Concrete,Transport and Road Research Laboratory supplementary report 748.

Neville, A M (1970) Creep of Concrete: Plain Reinforced and Prestressed, North-Holland, Amsterdam.

Neville A M, Dilger W H and Brooks J J (1983) Creep of Plain and StructuralConcrete, Construction Press, Longman Group, London.

Owen M J (1974) ‘Fatigue of carbon fibre reinforced plastics’, in Composite Materi-als, 5, ed. L J Broutman, London, Academic Press.

Plevris N and Triantafillou T C (1994) ‘Time-dependent behavior of RC membersstrengthened with FRP laminates’, J Struct Eng 120(3) 1016–1042.

Ponsot B, Valentin D and Bunsell A R (1989) ‘The effects of time, temperature andstress on the long-term behaviour of CFRP’, Composites Sci Technol 35 75–94.

Renton W J and Vinson J R (1975) ‘Fatigue behaviour of bonded joints in compos-ite material structures’, J Aircraft 12(5) 442–447.

Schaffer and Adams (1981) J. Appl. Mech. 48 859.Scott D W, Lai J S and Zureick A H (1995) ‘Creep behavior of fiber-reinforced

polymeric composites: a review of the technical literature’, J Reinforced PlasticsCompos 14(6) 588.

Shijie W and Ruixian Z (1993) ‘The study of fiber composite plates for strengthen-ing reinforced bridges’, Proceedings 9th Int. Conf. Composite Materials, Madrid,July 1993, pp 224–231.

Sillwood, J M (1982) Factors Concerning the Use of GRP in Corrosive Environ-ments, NPL Report DMA(B)20, Jan. 1982.

Standing Committee on Structural Safety (1983) ‘5th Report’, The Struct Eng61A(2) 65–69.

Swamy R N, Hobbs B and Roberts M (1995) ‘Structural behaviour of externallybonded, steel plated RC beams after long-term exposure’, The Struct Eng 75(16)255–261.

Taljsten B (1994) Plate Bonding – Strengthening of Existing Concrete Structures withEpoxy Bonded Plates of Steel or Fibre Reinforced Plastics, Doctoral Thesis, LuleaUniversity of Technology.

Tilly G P (1979) ‘Fatigue of steel reinforcement bars in concrete: a review’, inFatigue of Engineering Materials, 2(3) 251–268.

Tilly G P and Moss D S (1982) ‘Long endurance fatigue of steel reinforcement’,Proc. IABSE Coll. Lausanne, pp 229–238.

Tuttle M E and Brinson H F (1986) ‘Prediction of the long-term creep complianceof general composite laminates’, Exper Mech 26(1) 89–102.

Wake W C, Allen K W and Dean S M (1979) ‘The expected service life of structuraladhesives’, in Elastomers Criteria for Engineering Design, ed. C Hepburn andR J W Reynolds, Applied Science.

Wu, E M and Ruhmaan, D C (1975) ASTM Special Technical Publication 580, p 263.

SOR7 12/19/98, 3:41 PM221


222

8Analytical and numerical solutions

to structural strengthening of beams byplate bonding

P S L U K E

8.1 Introduction

One of the main successes of ROBUST has been the development ofpractical design rules and guidelines for flexural strengthening of reinforced(RC) and prestressed (PC) concrete beams. In general, designing a struc-tural member using a known code or guideline involves solution of ‘simple’design equations. Often to complement these equations, codes provideadditional data in the form of graphs, charts and/or tables assisting theengineer to obtain an optimum design. The additional data is normallygenerated experimentally or mathematically using closed form solutions ora suitable numerical technique. Experience has shown that exhaustive test-ing is a very expensive and time-consuming process and in recent yearsmore emphasis has been placed on numerical simulation to complementtesting. The development of high speed computers and more sophisticatednon-linear constitutive material models capable of simulating exactly whathappens experimentally has helped to make this transition. Previous chap-ters have described in detail the wide range of experimental tests that havebeen carried out in the laboratory and on site. These tests covered a widerange of parameters, but it is possible to use analytical techniques to extendthe parameters to enable additional data for the development of designrules for use in the construction industry.

The role of the finite element (FE) method in predicting the measuredexperimental results carried out in the project is discussed. Its reliability ingenerating design data, thus forming part of the process of developingpractical design rules and guidelines for use of these new materials in theconstruction industry, is also discussed. Some of the problems associatedwith prediction of non-linear behaviour of cracked reinforced and pre-stressed concrete beams using the finite element method are examined. Theresults obtained from finite element analysis and experimental tests arecompared and a series of recommendations are made.

SOR8 12/19/98, 3:46 PM222

Analytical and numerical solutions 223

8.2 Classical analysis

Researchers (Meier, 1987; Deuring, 1993) have stated that with the substi-tution of appropriate properties, classical reinforced concrete beam theorycan be applied safely when using advanced composite plates to strengthenbeams in flexure. Strengthened full scale T-beams tested at EMPA(Deuring, 1993) also seem to validate the strain compatibility method usedin the analysis of cross-sections. These statements have been examinedcarefully in the light of results obtained from beam tests in ROBUST. Anextensive desk study using the classical theory has been carried out on theROBUST test beams to predict the load/deflection history, the deflection ofthe beams and the mode of failure of the beams. Classical theory assumesthat plane sections remain plane throughout the load history.

The classical theory in general predicted the load/deflection curves rea-sonably accurately bearing in mind the inherent assumptions of the method.However in most cases, while classical theory predicted that the plateswould fail in tension, this phenomenon was not experienced in any of thetests on beams with unstressed plates. In addition, the concrete compressivestrain at failure in the test beams varied from 0.19–0.25%, whereas a strainof 0.35% is normally used for design purposes. It can be deduced from thetest results that due to the inherent conservatism built into the classicaltheory, it can be used safely to design beams strengthened in this manner.

8.3 Finite element analysis

Most classical theories for reinforced and prestressed concrete design arebased on relatively simple equations. These theories are generally based onlinear elastic models. They are usually not capable of dealing with problemswhere gross material and geometric non-linearities exist. In the absence ofdesign formulae to model accurately the behaviour of reinforced and pre-stressed concrete beams strengthened with advanced composite materials,the finite element method has been used to model the experimental data.The finite element method is a versatile and powerful numerical techniquefor analysis of engineering problems. Being a numerical tool, accuracy ofthe generated results is dependent on a number of factors.

One of the main approximations associated with non-linear behaviour ofconcrete problems is accurate modelling of concrete cracking. Under appli-cation of load, concrete cracks in the tension zone when stresses exceed itstensile strength. The number of cracks and their propagation pattern acrossthe depth of a concrete beam are unique and impossible to predict accu-rately. In addition, when concrete cracks, the stress path becomes discon-tinuous and the load transfer changes at the cracked section. Developmentof accurate material criteria and concrete elements that would model dis-

SOR8 12/19/98, 3:46 PM223


crete cracking of concrete is the most challenging part of any numericaltechnique.

The finite element technique has evolved over the past few years andtoday many sophisticated constitutive models are available which can rea-sonably predict various features of concrete behaviour including cracking,crushing and creep. Various procedures have been adopted for predictingcracking in concrete and these fall broadly into two categories, namely thediscrete crack approach and the smeared crack approach.

The smeared crack approach has been adopted widely for predictingthe non-linear behaviour of concrete (William and Warnke, 1975; Rotsand Blaauwedraad, 1989). In this approach cracks are modelled as localdiscontinuities which are distributed or smeared within the finite elementmodel. Using this technique, relative displacements of crack surfaces arerepresented by crack strains and the constitutive behaviour of crackedconcrete can be modelled in terms of a stress–strain relationship. Thisapproach fits the nature of the finite element method, as the displacementcontinuity field remains intact. However, it does have shortcomings, as ittends to spread crack formation over the entire structure, which makes itdifficult to predict localised failures. This has not stopped the method beingused widely and accurately for predicting load/deflection, load/strain be-haviour of concrete.

The discrete crack approach seems the most obvious way of modellingconcrete, because of the nature of concrete cracking which lends itself to adiscontinuous nature. With this technique, cracks are introduced into thefinite element model using interface elements between the concrete ones.The problem with this approach is that the position and direction of crackgrowth within the model is predefined, which is not so for concrete. Numer-ous models are being developed to introduce automatic remeshing tech-niques, which would help in alleviating some of these problems and allowcracks to propagate and change direction throughout the model.

Most commercially available finite element codes have material modelscapable of modelling concrete. Codes such as ANSYS, ABAQUS andLUSAS have models which treat concrete when it cracks as a smeared bandof cracks. A more recent finite element code, DIANA, has developedconstitutive models for concrete whereby the cracks are treated as discretecracks. Damage models have also been developed in recent years for mod-elling non-linear effects in materials. The approach has been extended toconcrete and offers a powerful alternative to the traditional smeared anddiscrete crack methods. It is quite clear from the developments in non-linear material models that more versatile models will be available in themajor finite element codes in the next few years.

Within the ROBUST project ANSYS, ABAQUS and LUSAS were usedto model the beams tested in the laboratory and on site. Oxford Brookes

SOR8 12/19/98, 3:46 PM224


University and University of Surrey used LUSAS and ABAQUS respec-tively to model their beams. This chapter mainly discusses the work carriedout by Mouchel in developing numerical models to validate and predict theexperimental results, both in the laboratory and on site.

Mouchel’s main finite element package (ANSYS, 1993, 1994, 1995) wasused to carry out analysis of over one hundred 1.0 m, 2.3m and 4.5 mreinforced concrete beams tested in four point bending under the ROBUSTproject at both Oxford Brookes University and University of Surrey andten post-tensioned prestressed concrete beams removed from the A34Botley flyover in Oxfordshire, UK and tested on site by the Royal MilitaryCollege of Science. Initially, the analysis was carried out using eight nodedquadrilateral isoparametric two-dimensional (2D) plane–stress elementsand later using eight noded three-dimensional brick elements to model theconcrete.

8.3.1 Two-dimensional analysis

In the 2D analysis, two plasticity material options were utilised to modelcracking of concrete, namely Drucker-Prager (DP) failure criteria andbilinear kinematic hardening (BKIN). Material characterisation tests werecarried out in the laboratory to determine properties of the steel rebar,carbon fibre reinforced polymer (CFRP) and glass fibre reinforced polymer(GFRP) plates and the concrete used in the analysis, and these were usedas a starting point for the 2D analyses. Typical measured values of themodulus of elasticity (E) for the beams tested in the laboratory were35 GPa, 205GPa and 140 GPa for the concrete, steel and CFRP, respec-tively. The measured compressive strength (fcu) for the concrete rangedfrom 46–60 MPa with its tensile capacity measured at 1.5–3MPa. Mild steelbars were used for the 1.0m beams with a yield stress of 350 MPa and anultimate stress of 460 MPa. High yield steel bars were used for the 2.3m andthe 4.5 m beams with a measured yield stress of 460MPa and an ultimatestress of 556 MPa. CFRP is a brittle material with a linear–elastic stress–strain relationship to failure. The measured ultimate tensile capacity was1500 MPa.

Although the behaviour of the plated beams was modelled with accept-able accuracy using DP, satisfactory results were not obtained when at-tempting to trace the load deflection path of the unplated beams. DP usesthe Von-Mises yield criterion with the associative flow rule. The materialresponse here is elastic perfectly plastic. The yield surface does not changewith progressive yielding, hence there is no hardening rule. Input param-eters required are an angle of internal friction and dilatancy and a cohesionvalue. Research has been carried out to determine these input parametersfor concrete, but they can be variable. To cater for this variability, a major

SOR8 12/19/98, 3:46 PM225


parameter study was carried out to study the sensitivity to the input values.It was observed that the models were very sensitive to these values. Inaddition, the FE models had to be fine tuned by adjusting material proper-ties for every test, which rendered its results unreliable for predicting be-haviour of other beams in the absence of experimental data. The secondmaterial option examined was BKIN, which uses the Von-Mises yield crite-rion with the associated flow rule and kinematic hardening. This is bettersuited for metals as failure of concrete is brittle unlike most metals, whichare ductile. Other material options available in ANSYS, namely themultilinear and anisotropic failure criteria, were also examined with theBKIN giving the closest correlation with the experimental results, and thiswas adopted in the 2D analysis without the hardening option.

The BKIN option in ANSYS assumes that the material strength in bothtension and compression is equal. Concrete has a different strength intension and compression hence it was necessary to divide the mesh gener-ated for a beam into regions with different material properties as shown inFig. 8.1. Regions ‘1’ and ‘2’ show the initially assumed tension and compres-sion areas of the beam. This was later modified by varying the tension depthof the beam after examining results of the analysis to find failure either inthe concrete or the CFRP plate. This was repeated until failure was ob-tained. Figure 8.2 shows that the load deflection graph generated by the FEanalysis is reasonably close to the experimental results. The FE strainvalues of the CFRP plate did not agree well with the measured experimen-tal results. In addition, the method was rather time consuming as it requiredseveral iterations to modify the tension–compression regions to achievefailure. In these FE analyses the actual measured material properties wereused throughout the iterations.

In general, the 2D analyses gave an indication of the behaviour of thestrengthened beams, but as the constitutive models were based on plasticitymethods more akin to metals, the results obtained could not have been usedwith confidence to predict the behaviour of the beams.

Figure 8.1 Division of 2.3m beam for two-dimensional analysis.

SOR8 12/19/98, 3:46 PM226


Figure 8.2 Load deflection plot for two-dimensional analysis of a 2.3mlong concrete beam.

8.3.2 Three-dimensional analysis

A three-dimensional brick element (SOLID65) in ANSYS, called ‘3D Re-inforced Concrete Solid’, is generally used for modelling solids with orwithout reinforcing rebars. In concrete applications, the element has thecapability to model the concrete while the rebar capability is available formodelling the reinforcement behaviour. The most important aspect of thiselement is the treatment of non-linear material properties. The concreteis capable of cracking, crushing, plastic deformation and creep in threeorthogonal directions at each integration point. When cracking occurs at anintegration point, the cracking is modelled through an adjustment of mate-rial properties which effectively treats the cracking as a ‘smeared band’ ofcracks, rather than discrete cracks. The presence of a crack is representedthrough modification of the stress–strain relations by introducing a plane ofweakness in a direction normal to the crack face. A shear transfer coeffi-cient is also introduced which represents a shear strength reduction factorfor those subsequent loads which induce sliding (shear) across the crackface. There are 16 possible combinations of crack arrangement and appro-priate changes in stress–strain relationships incorporated in the element.The open or closed status of integration point cracking is based on a strain

SOR8 12/19/98, 3:46 PM227


value called the crack strain. If this value is less than zero, the associatedcrack is assumed to be closed. If the value is greater than zero, the associ-ated crack is assumed to be open. When cracking first occurs at an integra-tion point, the crack is assumed to be open for the next iteration.

If crushing occurs at an integration point, material strength is assumed tohave degraded and has no contribution to the stiffness of the element at theintegration point. A number of input parameters are required to define theconcrete material data in addition to the E and Poisson’s ratio values. Thesefactors include the tensile cracking stress, crushing stress and shear transfercoefficients for open and closed cracks. Parametric studies were carried outon the shear transfer coefficients to test their sensitivity to the analysis. Itwas generally found that coefficients of 0.8–0.9 seemed to give good resultsfor all the modelled beams.

Three-dimensional tension or link elements were used to model thereinforcement explicitly, as the reinforcement option of the element uses asmeared approach. Eight noded plate elements were used to model theadhesive and CFRP. Bilinear and multilinear material failure criteria wereused for the CFRP and reinforcement, respectively. Figure 8.3 shows atypical finite element mesh of a 2.3m beam tested at Oxford BrookesUniversity (due to symmetry, one quarter of the beam was modelled foranalysis). Figure 8.4 shows comparison of results of the FE analysis with theexperimental values. It can be seen that the FE results compared favourablywith the experimentally measured values for both plated and unplated

Figure 8.3 Three-dimensional mesh of a 2.3m concrete beam (quartermodel).

SOR8 12/19/98, 3:46 PM228


Figure 8.4 Load deflection plot for three-dimensional analysis of a2.3m long concrete beam.

concrete beams. The strain values of the CFRP plate calculated by FE alsoagreed well with the experimental results as shown in Fig. 8.5. Actualmeasured material properties were used throughout the load history of thebeams, giving credence to the results.

This method of modelling the beams using three-dimensional finite ele-ment techniques has also been used to provide an accurate prediction of theload–deflection and load–strain history of the experimentally tested beamsprior to carrying out the experimental tests. Figure 8.6 shows load deflec-tion curves of 1.0 m beams tested at University of Surrey. The FE resultswere generated prior to testing the beams.

8.3.3 Effect of mesh density

When undertaking any finite element analysis, and in particular a non-linear analysis, it is extremely important to choose suitable elements and anassociated mesh in order to obtain a satisfactory solution to the problem. Asthe method is approximate, it is important to have a good understandingof the consequences of the assumptions when choosing the element typesused and size of mesh. This allows the effects of the approximation to be

SOR8 12/19/98, 3:46 PM229


Figure 8.5 Plot of experimental and FE plate strains for 2.3m longconcrete beam.

Figure 8.6 Load deflection graph for three-dimensional analysis of1.0m long concrete beam.

minimised within the solution. It requires skill and experience to be able todefine a mesh that will produce accurate answers economically for all typesof structural systems. In effect, the art of using the finite element methodlies in choosing the correct mesh density required to solve a problem. If themesh is too coarse then the inherent element approximations will not allow

SOR8 12/19/98, 3:46 PM230


an accurate solution to be obtained and alternatively, if the mesh is too finethe cost of the analysis can be out of proportion to the results obtained.In order to address this problem, a linear and non-linear parametric studywas carried out on the beams with meshes ranging from a fairly coarse to avery fine size. The results indicated that for a coarse mesh in particular, thenon-linear analysis solutions were failing to converge and the results ob-tained were in some cases 50% different from the actual results. However,the solution results obtained from an economic mesh size and comparedwith one that was ten times more dense, were shown to be within 1% ofeach other; the solution time for the latter problem was more than threetimes that for the former. The problem utilising the fine mesh above re-quired 7 h solution time on a Pentium 90 with 32Mb of memory. It is clearlyimportant, therefore, to choose a suitable mesh density for an analysis, toenable a balance to be obtained between the operation time and the accu-racy constraints. Once this exercise had been undertaken for one set ofbeams, the resulting mesh could be used satisfactorily with minimal modi-fications made to it on other beam sizes within the programme.

8.4 Effect of adhesive material

A parametric study was carried out to determine the effect of the adhesivelayer on the analysis and whether this layer could be ignored in the numeri-cal analysis. The results of the analysis showed that peel and shear stressesincreased non-linearly with an increase in adhesive elastic modulus. How-ever, the predicted peak peel and shear stresses at the adhesive/concreteinterface were low compared to the tensile and shear strengths of theSikadur 31 adhesive used in the ROBUST project. This confirms the testresults showing the concrete to be the weak link in the concrete adhesivelayer. No significant change in stresses in the plate and load/deflectionbehaviour was observed as a result of omitting the adhesive layer. In addi-tion, stress distributions at the concrete adhesive interface and at the con-crete/external plate interface (for the model with no adhesive) showedsimilar characteristics. Consequently, it was decided not to include theadhesive layer in subsequent analyses, as it had little effect on the results,but had a major effect on solution times and mesh sizes.

8.5 Prestressed 18.0m concrete beams

As a result of the success of the FE models to predict accurately the load–deflection and load–strain history of the 1.0 m, 2.3m and 4.5 m beams testedin the laboratory, the models were further developed to predict the behav-iour of the 18.0 m post-tensioned prestressed concrete beams removed fromthe A34 Botley flyover in Oxford which were tested in four point bending

SOR8 12/19/98, 3:46 PM231


on site at Kidlington as described in Chapters 4 and 5. The ten beams, whichwere 18.0m long by 711 mm deep by 343mm wide, were originally stressedwith five tendons. The original bridge had a reinforced concrete deck slabon top of the beams. As the beams in the ROBUST test conditions hadtheir top slabs removed, the five tendons in the case of the beams wouldhave represented an over-reinforced section. Hence three of the tendonswere cored at 1.5m centres in four places within the middle-third sectionof the beams, thus weakening the beams and making them suitable forstrengthening by plate bonding. Furthermore, as there were only tensuitable beams on which to perform the site trials, it was necessary tochoose parameters that would give the most meaningful results. The successof the finite element analysis in predicting the behaviour of the beamstested in the laboratory was a major factor in using the method to design thetests to be carried out on the site beams. The key parameters investigatedincluded the length of plate, thickness of plate, number of layers of plates,end anchorage conditions of plate, the effect of plate prestress and stresstransfer from the internal reinforcement to the external plate during coring.Core samples of the concrete were taken and tested to provide informationon the strength and stiffness of the concrete as input data for the FEanalysis. The measured E value was 35GPa and fcu was 42MPa. The yieldstress of the prestressing tendons was 1090MPa and the ultimate stress was1300MPa.

Figure 8.7 shows the finite element mesh used to model the 18.0m beams.As in the laboratory tests, the beams were simply supported on concreteabutments and loaded in four point bending. As the beams were symmetri-cal about their length and their cross-section, only one-quarter of the beamwas modelled, reducing computing time and resources. Advanced featuresof the ANSYS program were used successfully to model the effects ofprestressing.

Finite element analyses were carried out on models having two and threetendons, respectively. Although three of the five tendons had been cored inthe middle sections of the beams, prestressing strands on either side of thecore holes were still grouted and there was a possibility of some of the coredtendons reanchoring. The tendons outside the cored areas would then stillbe under a state of stress. It was anticipated that two tendons would give alower bound solution and three tendons an upper bound solution. Figure8.8 shows load–deflection plots for six unplated beams tested on site and thecorresponding FE predictions for the beams with two and three tendons,respectively. Figure 8.8 also shows clearly that the FE predictions effec-tively bound the experimental results. The results also show a small varia-tion in stiffness of the test beams. This figure shows very good correlationbetween the numerical and experimental results, lending further credenceto the use of calibrated FE models as predictive tools.

SOR8 12/19/98, 3:46 PM232


Figure 8.7 Three-dimensional mesh of an 18.0m prestressed concretebeam (quarter model).

Figure 8.8 Experimental and numerical load deflection plots forunplated 18.0m beams.

SOR8 12/19/98, 3:47 PM233


8.6 Beams with unstressed plates

Eight of the ten beams were strengthened with unstressed composite plates;beam numbers 1, 3, 4, 5 and 10 had three 90mm wide by 1mm thick plates,with various lengths and end conditions. Beam 1 had plates 12.5m long andunanchored at the ends. Beams 3 and 5 had plates 6.0m long; the ends ofbeam 3 were unanchored and those of beam 5 were anchored. The ends ofbeams 4 and 10 were unanchored and had composite plate lengths of 15.8mand 9.25m, respectively. FE analysis had revealed that peeling would notoccur at the ends, even on the 6.0 m plates. Figure 8.9 shows load–deflectioncurves for beams 1, 3, 4, 5 and 10 and the corresponding FE predictions. TheFE results for the variable length plates and the anchored plates showedlittle difference in strength and stiffness, hence only the worst response ispresented here. This phenomenon is also borne out in the experimentalresults. An 18.0m beam bonded with 6.0m long unanchored plates providesapproximately the same increase in stiffness and strength as beams havingunanchored plates, 12.5m or 15.8 m long, and as a beam with 6.0m longanchored plates; all these beams had a single layer of three 90mm wide by1mm thick plates side by side. However, stresses at the ends of the platesare higher in the shorter plates than in the longer plates, but not signifi-cantly so; a result predicted by the finite element analysis.

Figure 8.9 Experimental and numerical load deflection plots forbeams with one layer of 1.0mm thick plates.

SOR8 12/19/98, 3:47 PM234


Again, it is clear from the results that the numerical and experimentalsolutions are in close agreement, with the numerical results effectivelybounding the experimental ones, with the exception of beam 10, whichshows a stiffer response than the upper bound FE curve. This can beattributed to the fact that this beam was plated prior to coring and testing.Stresses from the cored tendons were transferred to the plate, thuspretensioning the plate. As a result the formation of cracks was delayedduring testing, due to the prestress in the plate resulting in a stiffer initialresponse.

Experiments in the laboratory on smaller scale beams had highlighted theprobability of peeling at the ends of unanchored short plates. This was notborne out in the large scale beams either experimentally or numerically,however beam 5 (6.0 m long by 1mm composite plates) was anchored at itsends to cater for this possible eventuality. Plate strains at the ends of theunanchored 6.0 m long plated beam were much higher than the correspond-ing anchored beam, as the bolts were taking much of the shear force withinthis region. However, the unanchored plate did not peel off and thestrengthening achieved was comparable. This end peel phenomenon on thebeams tested in the laboratory could be due to scale effects, with the beamsnot having sufficient anchorage length in which to dissipate the large strainsin the plate at the positions of cracks.

It was evident from all the tests carried out that coring significantlyweakened the beams locally in addition to the removal of the tendons. Allthe beams which were cored prior to plating had vertical cracks at the corelocations when loaded. Strain gauges were attached to the plates directlybelow the core holes and also midway between them. Figure 8.10 showsplots of plate strain versus distance from the plate for the numerical andexperimental results for beam 5. It is clearly evident that at core locations,where discrete cracks are clearly visible, the strain is very high, droppingdown to a much lower value between the cores. The figure shows two curvesfor the FE plots, one with the discrete cracks introduced directly into themodel from site measurements and the other showing the original FEmodel with smeared cracks. The FE model with discrete cracks gives goodcorrelation with the experimental results, but the FE model without thecracks appears to give reasonable average values for the stresses in the plateat various locations.

Beams 2 and 6 had two layers of plates 1 mm thick bonded to their soffits,the ends of beam 2 being unanchored and those of beam 6 being anchored.Beam 2 had composite plates 15.8 m long, whereas beam 6 had plates 6.0mlong. Figure 8.11 shows the load–deflection plots of the numerical andexperimental results for these two plated beams. The numerical results forthe two and three tendons of the beams are shown to bound beam 2, whichwas to be expected; this beam was cored prior to plating. Beam 6, which was

SOR8 12/19/98, 3:47 PM235


Figure 8.10 Plot of experimental and FE plate strains for beam no 5

Figure 8.11 Experimental and numerical load deflection plots forbeams with two layers of 1.0mm thick plates

plated prior to coring, shows a similar type of response to that of beam 5with regard to stiffness (Fig. 8.9); it shows clearly that the beams with shortplates gave similar increases in strength and stiffness to beams with longplates. Figure 8.12 shows numerical and experimental load–deflection plotsfor beam 9, which was strengthened with three layers of 1mm plates, 15.8mlong.

SOR8 12/19/98, 3:47 PM236


Figure 8.12 Experimental and numerical load deflection plots forbeams with three layers of 1.0mm thick plates.

Numerical analysis predictions gave very good correlation with the ex-perimental results in all the beams strengthened with unstressed plates.Results also demonstrated that end anchorages were not required for any ofthe beams, another feature predicted by the FE analysis.

The majority of the beams were tested to 80% of their predicted ultimatecapacity with two tendons intact, when unplated and subsequently whenplated. Extrapolated results show that a single layer of 1mm thick platesstrengthened the beams by between 23–26%. The FE results predict astrengthening of 21–23%, much in line with the experimental results. Thislevel of strengthening is sufficient to upgrade bridges from their current 38tonnes load limit to the 1999 EC requirement of 40 tonnes.

FE and experimental predictions for beams with two layers of 1mmplates showed increased strengthening in excess of 60% for the plated beamover the unplated beam, sufficient to strengthen bridges currently rated at25 tonnes to the 40 tonnes EC directive. Beam 2 was actually loaded veryclose to failure, and failure was due to triangular flexural cracks, whichcaused local debonding of the plate over a 200mm length near a core hole.Beam 9, which had three layers of composite plates bonded to its soffit, wasalso loaded very close to failure, with a similar type failure to beam 2, withan increase in strength of over 80%.

8.7 Beams with stressed plates

A method of stressing the plates was developed in the laboratory and issuitable for small span beams up to 8.0 m. The ROBUST project utilised

SOR8 12/19/98, 3:47 PM237


experimental laboratory beams ranging from 1.0–4.5 m (see Chapter 5). Inwork undertaken by the University of Surrey, a number of 1.0m, 2.3m and4.5m beams were tested under four point loading in the laboratory withCFRP plates that were prestressed to 25%, 40% and 50% of their ultimatetensile capacity.

The following observations by the ROBUST team on beams prestressedwith composite plates are:

1 The beams with plate prestress were clearly much stiffer than equivalentunstressed plates.

2 The action of stressing the plates is to reduce the onset of cracking,resulting in a stiffer member.

3 The in-service durability is enhanced due to reduced cracking.4 Prestressing the plates can close cracks on structures.5 The plate can be used as a safety net or reserve of strength against

failure of tendons in prestressed beams as a result of corrosion.6 The prestress also contributes to enhancing the shear capacity of the

beam.

In summary, prestressing CFRP plates can have quite a beneficial effecton extending the service life of existing bridges and increasing their loadcarrying capacity in shear and bending significantly.

As the technique was shown to be viable in the laboratory, a purposebuilt prestressing device was designed and manufactured under theROBUST project to be used to stress plates and bond them onto two of theten 18.0m prestressed beams on site at Kidlington. The device was capableof stressing one or two plates at a time, by anchoring one end of the platesand pulling from the other end. As each beam had three plates bonded sideby side to the soffit, the sequence of operation was first to stress the middleplate, followed by stressing the two outer plates. The prestressing operationwas implemented on site on beams 7 and 8 with three 90mm wide by 1 mmthick plates stressed to 30% of their ultimate tensile capacity of 1500MPaand then bonded onto the soffits of the beams. Beam 7, where the plateswere 6.0m long, had been cored prior to stressing and bonding, whereasbeam 8, where the plates were 15.8m long, was cored following this process,so that the transfer of stress from the tendons into the plates could bemonitored subsequent to coring. One of the outer plates on beam 8 failedduring the stressing operation at around 300MPa. This was due to a manu-facturing defect in the plate. In the event the beam was tested with only twoplates.

Figure 8.13 shows load–deflection plots for beam 7 (6.0 m long compositeplate) together with the corresponding FE plots for the beams with two andthree tendons, respectively. The results suggest that this beam is initially

SOR8 12/19/98, 3:47 PM238


Figure 8.13 Experimental and numerical load deflection plots forbeams with one layer of 1.0mm thick prestressed plates.

stiffer than both the FE plots, but then tends towards the FE plot with twotendons. This could be attributed to the fact that as the prestressing deviceneeded to be bolted to the vertical sides of the beam, drilling to accommo-date the bolts damaged the two remaining intact tendons locally. Thiscaused the steel to yield prematurely, further weakening the beam at thiscore location; a triangular crack distribution developed as the beam wasloaded and this resulted in local debonding of the plate at this location. It isinteresting to note that at the non-stressing end where a similar core holeexisted, but with no bolting required for a stressing device, similar crackingdid not occur, lending further credence to the possibility of damage tothe uncored tendons. Figure 8.14 shows plots of stresses in the plate forthe FE and experimental results. As mentioned previously, the FE modelswere modified to take account of the discontinuities at the core locationsdue to coring. The experimental curve shows peak stresses at the corelocations, but reduced in value somewhat between cores. The core, whichregistered a peak stress of over 1000 MPa, was where the plate haddebonded locally.

Numerical analysis predictions are in line with the experimental results,with the discrete crack model giving a similar curve to the experimentalcurve, and the smeared crack model giving a curve averaging the stress atthese locations. This beam was tested very close to failure and strengthen-ing of over 40% was achieved. This compares with 26% for an unstressedplate, highlighting the benefits of prestressing in terms of stiffness andstrength.

SOR8 12/19/98, 3:47 PM239


Figure 8.14 Plot of experimental and FE plate strains for beam no. 7.

8.8 Concluding remarks

Finite element methods utilising three-dimensional brick elements with aconcrete material criterion have been used successfully in calibration of1.0m, 2.3m and 4.5 m beams which were tested in the laboratory and havebeen employed to design subsequent tests on 18.0m beams with encourag-ing results. Finite element methods utilising two-dimensional plasticitymethods in ANSYS have proved to be unreliable as a predictive tool.Three-dimensional finite element methods have played a key role in devel-oping design rules and guidelines for plate bonded solutions using advancedcomposite materials.

Accurate prediction of cracking of the concrete and yielding of steelreinforcement of a reinforced/prestressed concrete beam using the finiteelement method is not a simple task, but has been achieved satisfactorilyunder this programme.

Work done on small scale beams and on 18.0m long beams with compos-ite plates prestressed before bonding onto the beams has been shown to beviable both numerically and experimentally. Prestressing composite platesrepresents a significant contribution to the advancement of plate bondingpractice. There is scope for increasing the shear resistance of reinforcedconcrete beams by bonding stressed plates to their soffits. Reduced crack-ing can thus enhance durability in service.

SOR8 12/19/98, 3:47 PM240


8.9 References

ANSYS (1995, 1996, 1997) The Theory, Procedures and Elements Manuals, ANSYSInc, Houston, USA.

Deuring M (1993) Post-strengthening of Concrete Structures with Pre-stressedAdvanced Composites, Report No. 224, CH-8600, EMPA, Switzerland.

Meier U (1987) ‘Bridge repair with high performance composite materials’, J Mater.Technik 4 125–128.

Rots J G and Blaauwedraad J (1989) ‘Crack models for concrete: discrete orsmeared? Fixed, multi-directional or rotating?’, Heron 34(1).

William K J and Warnke E D (1975) ‘Constitutive model for the triaxial behaviourof concrete’, Proceedings International Association for Bridges and StructuralEngineering, ISMES, Bermago, Vol 19, pp 174–190.

8.10 Bibliography

Fashole-Luke P S (1996) ‘Strengthening RC and PC beams with advanced com-posite materials’, British Cement Association (BCA) Conference, Sheffield.

Hutchinson A R and Rahimi H (1993) ‘Behaviour of reinforced concrete beamswith externally bonded fibre reinforced plastics’, 5th International Conference onStructural Faults and Repair, Edinburgh, pp 221–228.

Hutchinson A R and Rahimi H (1995) ‘Flexural strengthening of concrete beamswith externally bonded CFRP and GFRP plates’, Structural Adhesives in Engi-neering IV, Institute of Materials, Bristol, pp 177–182.

Lane J S, Leeming M B and Fashole-Luke P S (1997a) ‘Testing of strengthenedreinforced and prestressed concrete beams’, Construct Repair J 10 10–13.

Lane J S, Leeming M B and Fashole-Luke P S (1997b) ‘Using advanced compositematerials in bridge strengthening: introducing project ROBUST’, The Struct EngJ UK.

Peshkam V and Leeming M B (1994) ‘Application of composites to strengtheningof bridges: Project ROBUST’, Proceedings 19th International BPF CompositesCongress, Birmingham, 22–23 November, British Plastics Federation.

Peshkam V, Leeming M B and Fashole-Luke P S (1995) ‘The role of finite elementanalysis in composite plate bonding’, Conference in Infrastructure Design andRehabilitation, NAFEMS Research Working Group, London.

SOR8 12/19/98, 3:47 PM241


242

9Design and specifications for FRP plate

bonding of beams

M B L E E M I N G A N D J J D A R B Y

9.1 Introduction

The culmination of the research, the experimental testing in the laboratory,the modelling by numerical methods and the full scale beam testing on site,is the practical application of the technique to real situations. The first stageof this is calculating the extent of the strengthening required and determin-ing the quantity of the plates to be bonded. The next aspect that concernsthe designer is the specification of how the work has to be carried out. Thepractical aspects of undertaking the work on site have been covered inChapter 10.

9.2 Practical design rules and guidelines

Throughout the design process the client has to be satisfied that the tech-nique is sound and will give him the service life required. For techniquesthat have stood the test of time, a competent designer applies codes anddesign rules that have been formulated by experts and have had the ap-proval of appropriate authorities. In the case of new techniques, however,authenticated design rules do not exist and an experienced designer mustinterpret the research carried out. It is necessary to understand the limitsof the research and if there are areas where detailed knowledge is lacking,a conservative approach is necessary.

One of the objectives of the ROBUST programme of research was thedevelopment of practical design rules and guidelines for flexural strength-ening of reinforced (RC) and prestressed (PC) concrete beams. These rulesand guidelines were to be based on the experimental research and numeri-cal verification carried out. The preliminary rules will need to be modifiedlater by the results of further research and by taking account of the longterm performance of actual structures. The accuracy of numerical verifica-tion has been demonstrated in Chapter 8 by comparing those results withthe experimental ones reported in Chapters 4 and 5. The numerical

SOR9 12/19/98, 4:02 PM242

Design and specifications for FRP plate bonding of beams 243

methods used were developed for parameter studies of plate geometry andmaterial properties and although the technique can be used for design it isgenerally considered by engineers to be expensive, time consuming andcomplex for practical design. Consequently, simpler methods based onempirical formulae which will be derived from the numerical and experi-mental results must be developed.

9.2.1 Failure modes

The first step in the formulation of design rules is to determine how thestrengthened beam will fail under certain loading conditions and then toapply empirical formulae to predict that ultimate limit state of failure. Themember may have to satisfy more than one ultimate failure mode. Theapplication of suitable factors of safety will ensure that the failure mode isnot reached in practice.

Three main failure modes have been identified in the literature andwithin the programme as follows (see Fig. 9.1):

• Flexural failure, modes 1, 2 and 3 in Fig. 9.1, occurring either in thecarbon fibre reinforced polymer (CFRP) plate as tension failure, yieldof the steel reinforcement in tension or in the concrete as a compressionfailure, the first being likely to occur when the beam is under-reinforcedand the latter when over-reinforced. Yield of the steel reinforcement islikely to occur before either the concrete or the CFRP plate fails butwhile this may contribute to the ultimate failure of the beam it is not theprime cause of failure.

Figure 9.1 Typical failure modes for strengthened beams. Failuremodes 1–8 on figure. Failure mode 8, adhesive failure at concrete/adhesive interface; failure mode 9, adhesive failure at adhesive/FRPplate interface; failure mode 10, interlaminar shear within FRP plate.

SOR9 12/19/98, 4:02 PM243


• End anchorage peel, modes 6 and 7 in Fig. 9.1. The abrupt termina-tion of the plate can result in a concentration of stresses, some normal tothe plate, which cause the plate to peel off towards the centre of thebeam.

• Peel at a shear crack, modes 4, 5 and 8 in Fig. 9.1. This is a complexmechanism where debonding may occur due to strain redistribution inthe plate at the crack and/or the formation of a step in the soffit of thebeam causing shear peel. The delamination can then propagate towardsthe end of the plate. Whether mode 5 or 8 occurs depends on theamount of the shear reinforcement in the unstrengthened beam.

There are a number of other possible but unlikely modes of failure whichhave been identified in the literature such as delamination of the compositeplate or in the glue line, but these have generally not been experienced inthe research to date as the strength of these materials is higher than that ofconcrete. This type of failure could happen in practice if the installation hasbeen less than perfect or there is a defect in the manufacture of the plate.

9.2.2 Failure in flexure

Meier (1987) states that with the substitution of appropriate properties,classical reinforced concrete beam theory can be applied when using ad-vanced composite plates to strengthen beams in flexure. Strengthenedfull scale reinforced concrete beams tested by Kaiser (1989) at EMPA alsoseem to validate the strain compatibility method in the analysis of cross-sections. These statements were examined carefully in the light of theresults from beam tests in the ROBUST programme. A major study wascarried out using normal reinforced concrete theory to predict the load/deflection history, the deflection of the beams and the modes of failure ofthe beams.

In flexure, classical theory assumes that plane sections remain plane, thatis, strain compatibility is assumed. Deflection of the beams can be calcu-lated using simple formulae. Tracing the load/deflection history using nor-mal reinforced concrete theory gives a reasonable fit to the experimentalvalues, bearing in mind the inherent assumptions in the method. However,in most cases, while classical theory predicts that the plate will fail intension, this phenomenon has not been experienced in any of the tests onbeams with unstressed plates. The concrete compressive strain is usuallylimited to 0.35% for normal design purposes, whereas the value in the testshas been somewhat lower, in the range of 0.19–0.25%. Figure 9.2 comparesthe actual strains measured on one of the 2.3m beams tested at OxfordBrookes University with the calculated strains at a load of 90% of the actualfailure load in the test.

SOR9 12/19/98, 4:02 PM244


The strains on the concrete in the experimental work at Oxford BrookesUniversity were measured by demountable mechanical (Demec) gauges onthe side of the beams. It was not therefore possible to obtain readings atfailure load. The correspondence between measured readings and thosecalculated becomes worse as the failure load is approached. This is probablydue to the extensive cracking in the concrete that occurs in the tensionflange and to the yield of the reinforcement. However, the prediction is onthe safe side and will be more accurate at working load.

With the use of the classical reinforced concrete design theory and simpleformulae for deflection, load deflection graphs can be drawn and the resultscompared with the experimental results. The correspondence is good withthe calculated deflection following the changes in the slope of the experi-mental results as the reinforcing steel yields. The use of classical reinforcedconcrete theory is adequate for the purpose and has been found to predictstresses in strengthened beams better than the unstrengthened beams.

To apply reinforced concrete theory to strengthen beams it is necessaryto know the ultimate strength of the plates and their modulus and use anappropriate material factor of safety. The ultimate strength of the platesand their modulus is obtained from the characterisation tests carried out onthe plates. A material factor of safety, γm, of 1.5 is proposed for the RO-BUST CFRP plates, based on properties derived from test specimens fromthe production of plates, assuming a pultruded plate fully cured at the

Figure 9.2 Comparison of measured strains on a 2.3m beam withcalculated strains from RC theory.

SOR9 12/19/98, 4:03 PM245


works, and operating temperatures not in excess of 50°C combined witha heat distortion temperature greater than 90 °C. With these parametersEurocomp (1996) would advise that γm should not be taken as less than 1.5.This value can be compared with γm for steel of 1.15 and for concrete 1.5.The calculations can be carried out at the ultimate limit state, balancing thecompressive and tensile forces which have been calculated from straincompatibility using appropriate modular ratios for the various materials.

9.2.3 End anchorage peel

The mechanism of end anchorage peel has been studied by a number ofresearchers: Jones et al. (1988, 1989), Swamy et al. (1989), Roberts (1989),Taljsten (1994), Jansze (1995), Jansze and Walraven (1996), Zhang et al.(1995), Raoof and Zhang (1996), Oehlers and Ahmed (1996). The theory ofJones et al. (1988) relates to relatively stiff steel plates and calculationswithin the programme have shown that stresses predicted by the theory aremuch reduced due to the thinner more flexible CFRP plates. These stressesare said to be due to:

1 The force resulting from making the plate conform to the curvature ofthe beam; this force is 1/370 th of an equivalent steel plate.

2 Plate curl due to the eccentricity of the plate force to the bondline; thisforce is 1/14 th of the equivalent steel plate.

3 Interface bond stress or longitudinal shear; this force is still importantbut is reduced due to the CFRP plate being thinner, since the closer theoutermost fibre is approached the lower the longitudinal shear stress.

Swamy et al. (1989) suggested a design method of doubling the longitudi-nal shear stress and using an ultimate interface shear strength of √2� thetensile strength of the concrete.

The theory developed by Roberts (1989) is based upon partial interactiontheory. Incorporating the results obtained from the 1m beams into thetheory, anchorage shear/peel stresses at failure in the order of 14–18MPaare predicted. These are clearly not sensible, being considerably greaterthan the tensile strength of concrete.

Täljsten (1994) uses fracture mechanics for his predictions. Again thedata obtained from the tests on 1m beams at the University of Surrey andalso from the 2.3m beams at Oxford Brookes University were input into thetheory. Anchorage shear/peel stress at failure in the order of 1.3–1.8MPaare predicted for the 1m beams and from 1.8–3.4MPa for the 2.3m beamswhich did not appear to fail by this mechanism. These results are morereasonable with respect to the expected shear or tensile strength of concretealthough a value of at least double would have been expected for the 1mbeams.

SOR9 12/19/98, 4:03 PM246


Zhang et al. (1995) and Raoof and Zhang (1996) present a theory basedon a series of cantilever teeth of plain concrete between cracks in the coverzone anchoring the plate and is limited by the tensile strength of the con-crete. The spacing of cracking is difficult to determine but a minimum andmaximum spacing is said to give an upper and lower bound solution whichbrackets the data.

A simplistic analysis was carried out by calculating an anchorage stress ina similar manner to the bond stress for a reinforcing bar by taking the forcein the plate at the point of maximum moment at failure and dividing it bythe length multiplied by the width of the plate beyond that point. Stressesfrom this exercise varied from 2.6–4.2MPa which are of an order that mightbe expected although there is still a wide variation. Breaking these figuresdown into groups depending on plate width gives the results in Table 9.1.

None of the above theories presents an adequate method for generaldesign. End peel stresses have been shown to be critical for thick andrelatively stiff steel plates but it is thought that they are not significant forthin and flexible composite plates. There is no doubt that peel stresses occurat the end of a plate that is curtailed at a point some distance from thesupport but none of the above theories fitted the experimental results fromthe ROBUST programme.

The calculations for anchorage length suggested by European thinkingare based on tests carried out when a CFRP plate is bonded to a concreteblock and a force applied to pull the plate from the concrete in a directionparallel to the bond line. The tests are often done as a double shear lap test.In these tests the peel is initiated not at the end of the plate but at thefront face of the concrete. The peeling mechanism is not therefore dueto end peeling stresses but to high longitudinal shear stress at the leadingedge of the concrete. This condition would be closer to the situation at theedges of a crack. The failure mode is considered to replicate FRP peelingoff at the outermost crack in the uncracked anchorage zone. Calculationsbased on fracture mechanics give good agreement with experimental re-sults. Usually when a plate peels under these conditions only a thin layer ofconcrete a millimetre or two in thickness adheres to the plate but in somecases its failure line may divert into the plate causing interlaminar platefailure.

Table 9.1 Anchorage stresses for the 1m beams

Plate width (mm) 90 65 45Plate thickness (mm) 0.5 0.7 1.0Maximum stress (MPa) 2.62 3.34 4.22Minimum stress (MPa) 2.53 3.09 3.99

SOR9 12/19/98, 4:03 PM247


Rostasy (1993) looked at this situation and cites work by Ranish (1982)who studied bond stresses in the anchorage length of steel plates bonded toconcrete blocks. It was found that a maximum value of bond stress (max τb1)was reached which equalled four times the mean surface tensile strength ofthe concrete established by a pull-off test. The necessary anchorage length(lbl) was found from equation [1]:

nec bl lu bll F E t b5 221 1 1

2 515 10. . .max . .( ) [1]

where lbl is between 500–1500 mm, Flu is the limiting plate force, El is themodulus of the plate (MPa), t1 is the thickness of the plate for valuesbetween 5–15mm and b1 is the width of the plate (about 150mm). Thisformula was introduced into early design procedures for CFRP plateswhere the thickness of the CFRP laminate was expressed as the thickness ofan equivalent steel laminate in the ratio of their respective strengths. Themax τbl was also given in a table with values according to Tausky (1993)based on the measured tensile strength of the concrete at the surface. Thistheory assumed a minimum bond length of 500mm and the force to beanchored was the tensile force in the CFRP laminate at the design level inthe location of the maximum moment.

Neubauer and Rostasy (1997) carried out 51 bond tests using CFRPplates bonded to concrete blocks where the bond length, plate width, platethickness and concrete cube strength were varied. It was found that anchor-age lengths of between 195–330mm were required to sustain a plate forcethat could not be exceeded and that longer anchorage lengths were noteffective. Fracture mechanics was used to analyse the results which wereexpressed in the formulae [2]–[4]:

l E t ft. ctm max . .� √( ) ( )1 1 2 [2]

T k b E t fck. b 1 ctm 0.5.max . . . .� √( )1 1 [3]

where 1.06. 2 1b 1k b b b� � � �√ ( )( ) (( )( )1 4001 [4]

where lt.max is the maximum bond length in mm, E1 is the modulus of theplate, t1 is the thickness of the plate in mm, fctm is the concrete surface tensilestrength in Nmm�2 which should not be taken as greater than 3Nmm�2,Tck.max is the maximum plate force that can be anchored in Newtons, b1 is theplate width in mm and b is the width of the beam soffit or the plate spacingin slabs in mm.

Chajes et al. (1993) first investigated adhesives with moduli from5–0.2 GPa with various types of surface preparation in similar tests.The adhesive with the lowest modulus failed in the adhesive while theothers failed in the concrete. Fusor together with a Chemglaze primer (anorganofunctional silane) gave the best results and was chosen as the adhe-

SOR9 12/19/98, 4:03 PM248


sive for further tests where three different concrete strengths were used.These few results were used to predict a shear stress failure in the concreteof 11.1 √fc psi, where fc is probably the cylinder strength although not stated(0.92 √fc in MPa). Further lap shear tests were then carried out using 50 mm,100 mm, 150 mm and 200mm bond lengths with a 36.4N mm�2 concrete.These tests showed that a maximum force of 12 kN could be anchored by a25 mm wide strip with a 95mm anchor length and that longer anchoragelengths did not increase the force that could be anchored. An average shearstress in the concrete was estimated at 4.945 Nmm�2. Chajes et al. (1993)therefore came to the same conclusion as Neubauer and Rostasy (1997)that there is a certain anchorage length that will sustain a plate force thatcannot be exceeded. Table 9.2 compares the results from these two papers.

Table 9.2 Comparison of anchor lengths

Neubauer and Rostasy (1997) Chajes et al. (1993)

Adhesive Sikadur 31 Fusar

Tensile 24.81 30.6strength(MPa)

Elongation 0.41 3.0(%)

Elastic 5.1721 1.584modulus(GPa)

CFRP CarboDur PrepregPlates

Tensile 2000–3000 1655strength(MPa)

Elongation 3–5 1.5(%)

Elastic 150–230 108.5modulus(GPa)

Results Concrete Plate Anchor Plate Concrete Plate Anchor Platestrength thickness length force strength thickness length force(MPa) (mm) (mm) (kN) (MPa) (mm) (mm) (kN)

25 1.2 22825 2.4 330

36.4 1 95.25 1255 1.2 19455 2.4 275

1 Figures from Chajes et al. (1993) for Sikadur 31.

SOR9 12/19/98, 4:03 PM249


Detachment of the plate was a common failure mechanism in the RO-BUST tests. However, the plate became detached so quickly that there wasalways doubt about whether the failure started at the end of the plate andpropagated towards the centre or started peeling at a crack and propagatedtowards the end of the plate. Even recording the point of failure with a highspeed video camera could not resolve the question. It is now thought thatpeel initiating from a shear crack is the more likely mechanism.

9.2.4 Peel at shear or at wide shear/yield crack

Where the shear span to effective depth of the beam is less than about 6, thenormal mode of failure of a reinforced concrete beam is usually due to theformation of a diagonal flexure–shear crack. This effect is illustrated in Fig.9.3 where the results of the tests on 1m long beams at the University ofSurrey are presented. The effect of the width of the plate is evident, thewider the plate the greater the ultimate moment. At low values of shearspan to depth ratio, effective anchorage of the ends of the plates alsoincreased the ultimate moment. The shear span to depth ratio is related to

Figure 9.3 Effect of shear span to depth ratio and plate width onultimate moment of beams strengthened with CFRP plates.

SOR9 12/19/98, 4:03 PM250


the longitudinal shear stress between the plate and the concrete but with thescatter of the results a lower bound longitudinal shear stress could not beassessed. Whether this reduction in ultimate strength was due to shear orend peel effects could not be ascertained. It is clear that where the shearspan to depth ratio is low, calculating the required amount in strengtheningby flexure alone is not sufficient and shear and end peel effects must beconsidered or the plate ends must be anchored adequately. At a point whena crack occurs at the end of the plate due to shear and bending, end shearpeel is likely to occur and the whole depth of the cover concrete togetherwith the plate will detach and peeling will occur at the level of thereinforcement.

If a shear crack in a beam strengthened with a CFRP plate approaches anangle of 45°, a step can occur in the soffit of the beam as the end of the beamrotates about the compression zone of the top flange (see mode of failure 4in Fig. 9.1). This failure mode was identified by Meier and Kaiser (1991).This mechanism can cause the plate to peel due to stresses normal to theplate and may be aggravated by stress redistribution across the crack.Triantafillou and Plevris (1992) address the problem and propose the rela-tionship P � λ(Gs.As � Gp.Ap) where P is the debonding load, G is theshear modulus, A the area of the main reinforcement and the plate, respec-tively and λ is a constant determined by experiment. From the three experi-ments they found that λ approximates to 0.011 using a value of 4.4GPa forthe shear modulus of the CFRP plates.

Many failures within the ROBUST programme are now thought to haveinitiated from a near vertical crack at a point about a beam depth away fromthe load point associated with triangular cracking of the concrete followedby peeling towards the end of the plate. The vertical crack occurred be-tween the shear link reinforcement. The beams were heavily reinforced inshear to avoid shear failure of the type described above. A possible mecha-nism for failure mode 8 in Fig. 9.1 is the formation of a wider crack at thepoint where the steel reinforcement yields, leading to debonding of theplate to redistribute the high strains across the crack which then propagatetowards the end of the plate. This mechanism may be aggravated by somesmall vertical movement at the soffit due to shear strain across the crack.The factors influencing the failure will be:

• shear transfer in the compression zone of the concrete,• shear transfer due to aggregate interlock in the crack below the neutral

axis,• dowel action of the steel reinforcement,• dowel action of the CFRP plate, although this must be small and should

be discounted,• a shear step in the soffit of the beam causing plate peel. This latter

SOR9 12/19/98, 4:03 PM251


mechanism is dependent on the angle of the crack and the extent of theshear strain across the crack,

• the formation of the wide crack due to yielding of the steelreinforcement.

The main objective of the programme was to study CFRP plate bondingin flexure but it is clear that with low shear span to effective depth ratiosused in some of the experiments, the mode of failure was a shear peelingcausing the removal of the layer of concrete cover to the internal tensilereinforcement. All the beam tests in the programme were carried out underfour point bending. This loading configuration gives constant shear in theshear span with a high bending moment at the loading point that diminishesto zero at the support. This form is commonly used in loading tests as itleads to simple analysis but is not typical of loading in practice whereuniformly distributed loads are more common and point loads, when theyoccur, are frequently moving such as in wheel loads. With uniformly distrib-uted loading there is no clear shear span, and anchorage and shear effectsmay be less pronounced. Shear effects are not well understood in reinforcedconcrete theory and the relative effects of shear transfer in sound concrete,aggregate interlock in cracked concrete and the dowel action of thereinforcement are difficult to quantify separately. Taylor (1974) gives thefollowing proportions of the shear force taken by the above actions: com-pression zone shear 20–40%, aggregate interlock 33–50% and dowel action15–25%. Regan (1993) gives a more recent overview of research into shearover the last 100 years. It is not thought that the dowel action of the CFRPplate is significant. The extent of a shear step in the soffit of the beam isdifficult to quantify and is also probably of little significance. The situationwill be on the safe side if the beam satisfies the normal code requirementsfor shear ignoring the contribution of the CFRP plate.

The stresses in the plates and in the steel reinforcement can be calculatedthroughout the length of a beam according to classical RC bending theoryon the basis that the stresses vary in accordance with the moment at anypoint. This is illustrated in Figs. 9.4 and 9.5 for the half spans of two beams,one with a single point load and the other with uniformly distributed load-ing. The case of the single point load is similar to that of the shear span onlyof a beam loaded in four point bending as is usual in experimental beamtests.

Longitudinal shear stress between the plate and the concrete will berelated to the rate of change of the force in the plates and hence to the stressin the plate for constant section. It can be seen from Fig. 9.4 for the case ofa single point load that the greatest slope on the curve of the stresses in theplate is towards the middle of the beam where the steel reinforcement hasyielded and can no longer contribute any greater force to balance the

SOR9 12/19/98, 4:03 PM252


Figure 9.4 Tensile stresses under a single point load.

Figure 9.5 Tensile stresses under uniformly distributed load.

SOR9 12/19/98, 4:03 PM253


increasing moment. The longitudinal shear stress, at this point, can beshown to be approximately equal to S/(by) where S is the shear force, b thewidth of the plate and y the distance from the middle of the concrete stressblock which is effectively the distance of the centroid of the plate from thetop of the beam less half the neutral axis distance. For the case of auniformly distributed load the greatest slope is between about 0.6 and 0.7 ofthe half span. The former case is a coincidence of high moment and shear;the latter of high moment but low shear.

Combined with the above, higher stresses will occur in the CFRP plate atthe location of the crack due to the increase in strain over the width of thecrack. This instantaneous high stress can only dissipate by debonding eitherside of the crack, possibly causing a triangle of concrete to detach from oneside of the crack. This phenomenon was found in the testing of the 18mbeams where the strains in the plates were found to be very high at thepoints where the prestressing cables had been cut by coring through the sideof the beam. These strains were found to be 1.25 to 1.5 times the averagestrains. The debond will release the peak stress to a point where the rate ofchange of stresses in the plate no longer exceeds the longitudinal shearstress that the concrete can withstand.

Once the length of the debond is sufficient for the difference in the forcesin the plate on either side of the debond to exceed the maximum longitudi-nal shear stress of the concrete, the debond will propagate further and willultimately reach the end of the plate. The three factors involved in failuremode, namely, high longitudinal shear stress at the interface between theplate and the concrete, stress redistribution in the plate over the crack andshear step effects, do not lend themselves to simple analysis for designpurposes. It is suggested, therefore, that in the design the strain in the plateis calculated at the point where the internal steel reinforcement starts toyield and the ultimate load for this failure mode should then be taken as thepoint of yield strain plus an additional limiting strain. The additional limit-ing strain has yet to be determined from experiment.

9.2.5 Limit states and partial safety factors relevant tocomposite plate bonding

The design of strengthened bridges should be considered for both theserviceability limit state (including the checking of stress limitations) andultimate limit state, in accordance with the relevant clauses of BS 5400: Part4 (1990), except where amended in this chapter. The appropriate loads andload factors should be taken from BS 5400: Part 2 (1990).

Partial safety factors to the characteristic strength of the CFRP plates ofγm � 1.5 for ultimate limit state (ULS) and γm � 1.00 for serviceability limitstate (SLS) should be used. Where the characteristic strengths of the exist-

SOR9 12/19/98, 4:03 PM254


ing concrete and reinforcement are not known characteristic strengths maybe derived from test evidence.

The ‘over-reinforcement’ of a concrete section can result in a brittlefailure. Sections to be strengthened should, therefore, be checked to ensurethat this does not occur, in particular by using the requirements of Clause5.3.2 of BS 5400: Part 4 (1990).

9.2.6 Plate dimensions and spacing

When determining the size and thickness of each individual plate forstrengthening a slab, the effective width of the section should not be greaterthan the width of the plate plus twice the distance of the plate to the neutralaxis of the section.

9.2.7 Fatigue and creep

Fatigue is discussed in Chapter 7, Part 2 ‘Fatigue Behaviour’. The stressrange for fatigue purposes in the existing steel reinforcement of reinforcedconcrete beams should not be increased when the beam is strengthenedby CFRP plate bonding. In prestressed concrete it should be ensured thatstress transfer due to creep in the concrete will not result in excessivecompressive forces being induced into the plates.

9.2.8 End and intermediate fixings

Intermediate fixings are not required as is the case with steel plate bondingas the plates are much lighter. The ‘grab’ of the adhesive is such that theplate remains in position after having been rolled into the adhesive layerwithout further support.

Fixings may be required at the ends of the plates to resist concentrationsof peel and shear stresses in this region, where these stresses exceed theproduct of the pull-off strength of the concrete and γm. The fixings should bedesigned as described above. In the ROBUST programme of research,15 mm thick �45° glass fibre reinforced polymer (GFRP) end tabs werespecially bonded to the ends of the plates and stainless steel bolts werepassed through both the plate and the end tab at suitable spacings and wereanchored at a depth into the concrete below the steel reinforcement. Thebolts should be supplied with large washers and tightened up to a predeter-mined torque to prevent crushing of the composite materials. It is notadequate to omit the end tabs and merely to drill through the compositeplates and thence to anchor them at their ends with bolts. Drilling holesthrough unsupported composites will sever the unidirectional fibres and theconcentrated compressive forces under the bolt head will weaken the plate

SOR9 12/19/98, 4:03 PM255


further; it is not possible for the forces in the plate to be transmitted into thebolt. The bond between the ROBUST end tabs and the plates was designedto take the whole force to be anchored. This force is then transferred intothe concrete by the fixing bolt. Other forms of fixing were not found to beso effective. Tests by Poulsen (1996) show that the ultimate anchorage forcecan be increased by up to three times by utilising three bolts, but increasingthe number of bolts beyond this value is not effective. The effect of bondingshort lengths of laminate across the end of the plate was also examined.One transverse laminate increased the ultimate force anchored by 75% butfurther transverse laminates gave little increase in anchorage force.

Plates should not extend into areas of compression without further veri-fication about their adequacy, as plate buckling may occur causing tensile/peel stresses in the adhesive normal to the plate. The maximum com-pressive force in the plate is limited by the maximum permitted strain inthe concrete at compressive failure, normally taken for design purposes as0.0035. Where the plates are fully bonded they can be considered fullyrestrained against buckling. The out-of-plane forces can be calculated andeven an out-of-straightness of 5mm in 1m produces out-of-plane forces thatare well within the tensile strength of the concrete which is the weakest link.However, an unbonded length as short as 40mm for a 1.4mm thick platecould be liable to buckling failure. Clearly the length of voids in thebondline must be limited to well below the critical buckling length depend-ent on the thickness of the plate. During construction, very careful monitor-ing of voids in the bondline is required for sections of plate that will go intocompression during any part of the loading cycle. This is an area that wouldbenefit from further research.

9.2.9 Vandalism and fire

Where fire damage and vandalism are expected the plates may be coveredby a suitable polymer-modified cementitious screed or by suitable intumes-cent coatings. Fire tests have been carried out at EMPA where CFRP plateswere found to last considerably longer than steel plates. The reason for thisis that the adhesive is the most vulnerable factor and steel conducts the heatrapidly to the glue layer. Carbon does not conduct heat to the same extentas steel and to some extent protects the glue layer. Composite materials arebeing increasingly used in offshore structures as panelling for blast and fireprotection (Wu and Gibson, 1994).

9.2.10 Durability

At present there are no standard accelerated laboratory testing methods topredict the long term performance and durability of the system. The dura-bility of composite plate bonding is discussed in Chapter 6.

SOR9 12/19/98, 4:03 PM256


9.3 Application of the technique

9.3.1 General

The technique of bonding CFRP plates to structures can be used in alllocations where there is a requirement for additional flexural reinforcementin the tensile zone. Their use for strengthening in compression requiresfurther justification. Careful consideration should be given where the head-room of bridges is critical or where there is evidence of frequent damage tothe soffit from vehicles. Special provisions may be necessary to protect theCFRP plates under such circumstances. Where plating to soffits of bridgesabove carriageways is being considered the available headroom should bechecked. Due allowance should be made for any fixings required.

The Highways Agency (1994) requires that, for steel plate bonding, anyelement of a bridge structure to be strengthened should be capable ofsupporting nominal dead load, superimposed dead load factored by thepartial safety factor for loads, γfL � 1.2 and nominal HA (normal) live load(i.e. unfactored) when checked at ULS. There seems no reason to alter thisrequirement for CFRP plate bonding at present.

Where plates are bonded to the top surfaces of slabs and beams andsubsequently buried by the road surfacing it would be impractical to pro-vide inspection facilities for the plates. In such cases, special care should betaken during bridge inspections to identify any areas of plates that may havedebonded as indicated by local breakup or reflective cracking of the surfacein the location of the plates. It is essential that accurate drawings indicatingthe location of all plates are maintained and are readily available for suchinspections.

9.3.2 Investigations and tests

Before plate bonding is considered for a structure, investigations should becarried out to ensure that the risk of corrosion in the existing member is lowand that the structure is sound enough (including any repaired areas) forstrengthening by plating.

Surfaces that are damp or subject to leakage, particularly if contaminatedwith chlorides, should only be plated after satisfactory remedial measureshave been taken. If the reinforcement is corroding, the expansive rustproducts may disrupt the concrete and eventually cause debonding of theplate. Therefore, unless repairs have been carried out, plate bonding shouldonly be considered for members where chloride values are generally lessthan 0.3% by weight of cement and half-cell potential measurements arenumerically generally less than �350mV (e.g. �200 mV) with respect to acopper/copper sulphate electrode. CFRP plates are less vulnerable to dura-bility problems than steel plates and therefore surfaces that are damp and

SOR9 12/19/98, 4:03 PM257


subject to leakage are less critical. However, steps should be taken to rectifythe dampness and prevent the leakage. Tests in ROBUST carried out byOxford Brookes University showed that there was no loss in pull offstrength when Sikadur 31 was used to bond the dollies to saturated surfacedry concrete. Nonetheless, it is advisable to avoid these conditions to ensurethat the highest standards are used in installation of the technique. Whiledampness may not affect the composite plate and adhesion to the concretesignificantly, it could cause further corrosion of the steel reinforcementleading to spalling of the cover concrete to which the plate may be bondedand would also be susceptible to freeze/thaw damage causing delaminationof the concrete cover.

9.4 Materials

9.4.1 Composite plates

A variety of composite plates have been used in the ROBUST and otherresearch programmes, using glass, carbon or aramid fibres manufacturedby the pultrusion process or made up from prepreg material. In the UK andin Europe the use of plates containing about 60% by volume fraction ofunidirectional carbon fibre manufactured by the pultrusion process in anepoxy or a vinylester resin matrix is favoured. The plate in ROBUST was90mm wide by 1 mm thick and used T300 carbon fibre in a vinylesterresin (BASF A430) combined with a peel-ply inbuilt surface finish on bothsides of the plate. This method of surface finish provided a suitable surfacefor bonding without an extra manufacturing process and kept the surfaceto be bonded clean until just before the adhesive was applied. The choiceof T300 fibres was made on cost and availability at the time with a penaltyof a lower strength and modulus. T700 fibres are much more widely avail-able now and there is no reason why this superior strength of fibre shouldnot be used in plates of different widths and thickness. Other plates fromEurope have used T700 fibres and are available in various sizes but requireabrading and wiping clean before use to obtain a satisfactory bondingsurface.

The modulus of glass is too low to give satisfactory performance inplate bonding requiring at least three times the thickness of an equivalentCFRP plate to give the same stiffness and is therefore not being usedeconomically. Aramid has no special advantages over carbon and is moreexpensive.

The use of any plating material must have adequate data to support thequoted properties resulting from characterisation tests carried out in ac-cordance to appropriate standards such as BS2782 Part 3 (1996) and Part 8(1994), ASTM D3039/3039M (1995) and Crag (1988). These standards re-

SOR9 12/19/98, 4:03 PM258


quire a test sample 25 mm wide. Tests have been carried out on full widthsamples which gave significantly lower results. This size effect must beresolved before reliance can be placed on results from coupon specimens.The material factor of safety used in design should reflect this uncertaintyand will depend on the number of tests carried out and their variability.

9.4.2 Concrete – suitability for bonding

A range of tests is available to determine the suitability of the concretesurface for bonding plates. The most suitable test method involves bondinga metal dolly to the concrete using the same adhesive used to bond theplate, and subsequently pulling the dolly off. Examination of the failuremechanism and failure load gives useful guidance. The preparation is con-sidered suitable when a cohesive failure occurs in the concrete. This testshould be carried out in accordance with BS 1881 (1992) or with the latestedition of prEN 1542 (awaiting translation before ratification). A minimumvalue of 1.5 Nmm�2 should be obtained.

9.4.3 Adhesive

On the basis of long term experience epoxy resin adhesives have beenfound to be suitable for steel plate bonding. Their durability has beenestablished by use over a period of twenty years. The same epoxy resinadhesives have been used in research and in practice for CFRP plate bond-ing with acceptable results. An adhesive must demonstrate an acceptabletrack record of use in steel and CFRP plate bonding or must be subjectedto additional tests to prove its acceptability for strengthening proposalsdetailed below.

Epoxy resin adhesives require care in use. Manufacturers or formu-lators commonly supply two-part resins in containers suitably propor-tioned for mixing. It is important that all the hardener is added to theresin in its container and mixed with a slow speed mechanical mixer. Highspeed mixing entrains air and is less efficient. The resin and hardener shouldbe of different colours and adequately mixed to produce a uniform colour.The speed of the chemical reaction increases with the temperaturegenerated.

Sikadur 31PBA was used in the ROBUST programme of research. Thisis a two-part cold cured, epoxy-based structural adhesive composing a resinwhich is white in colour and a black coloured hardener. They should bemixed together to form a uniform grey colour. The adhesive should beapplied onto the CFRP and/or concrete surfaces within 20min. Sikadur 30has been used more extensively in Europe. The properties of Sikadur31PBA and Sikadur 30 are given in Table 9.3.

SOR9 12/19/98, 4:04 PM259


Table 9.3 Properties of typical adhesives used in plate bonding

Sikadur 31PBA Sikadur 30

Heat distortion temperature (°C) 43 62Flexural modulus (kNmm�2) 8.6 12.8Tensile strength (Nmm�2) 21.8Moisture resistance <0.5% after 28 daysDensity (kgm�3) 1500

9.5 Workmanship

9.5.1 Surface preparation of concrete surfaces

It is important that preparation of the concrete substrate is carried out wellto ensure an adequate bond with the adhesive. Minor protuberances andimperfections can be removed by scabbling or grinding the surface. Gritblasting is the preferred method of general surface preparation and shouldbe carefully carried out to provide a laitance and contamination free surfaceresembling coarse grained sandpaper with exposure of minor aggregate.Consideration should be given to filling any cracks or fissures (includingopen construction joints) wider than 0.2mm or liable to leakage, by injec-tion of a suitable, compatible low viscosity resin. The prepared surfaceshould be dust free and surface dry. If moisture is picked up by absorbentpaper pressed onto the concrete it is likely to be too damp for bonding. Itmay be necessary to provide temperature and humidity control to dry outthe concrete sufficiently prior to bonding.

9.5.2 CFRP surfaces

The ROBUST CFRP plate has an in-built peel-ply surface treatment in-stalled as part of the pultrusion process. The peel-ply provides a cleansurface, free from contaminants such as release agents, which is wettableand has an appropriate surface texture with a freshly fractured matrix resinsurface with a surface free energy value in excess of 55 mJm�2. The peel-plysacrificial surface layer is removed immediately before application of theadhesive. Plates which do not have this in-built peel-ply surface treatmentshould be abraded and cleaned and degreased with a suitable cleaningagent such as Sika Colma Cleaner. The plates should be maintained in a drycondition before application of the adhesive.

SOR9 12/19/98, 4:04 PM260


9.5.3 Application of adhesive

The bonding of CFRP plates to concrete members is carried out by applyingthe structural epoxy resin to the plate and/or concrete, offering the plate tothe soffit of the member and applying pressure to ensure intimate contactwith the concrete. It is important to spread the adhesive immediately aftermixing to dissipate the heat generated and extend its workability time.Common practice is to spread the adhesive slightly more thickly in a domedprofile along the centre line of the plate than at the sides of the plate. Thisreduces the risk of forming voids when pressing the plate loaded withadhesive against the concrete surface. Excess adhesive squeezed out of theplate can be scraped away and a fillet formed at the plate edge. Small glassbeads (ballotini) dropped onto the epoxy adhesive surface prior to fixing canbe used to maintain the minimum required adhesive thickness of 1–2 mm.Procedure trials should always be carried out to prove the method ofapplication and acquaint the operatives with the material. The ambienttemperature and dew point temperature should be determined and if theambient temperature is less than 10 °C or dew point is more than 10 °C,artificial heating and dehumidifiers may be required to maintain the ambienttemperature and humidity at acceptable levels for installation and curing.The manufacturer’s instructions on safe use of resins should be followed.

Based on experience in ROBUST up to three plates may be bondedin layers to give the required total thickness, each plate being bondedseparately.

9.6 Quality control

9.6.1 Adhesive bond/shear tests

In bonding to CFRP, assurance of the quality of adhesive bonded jointsshould take place prior to bonding by ensuring adequate control of all theprincipal steps in the bonding process; these steps include surface prepara-tion, mixing, application and curing of the adhesive. The purpose of the testis to check adhesion and the adhesive material properties. Simple single lapshear tests are recommended, with a 10mm overlap, see below. Threesingle lap shear tests should be performed for each batch delivered to thesite. The CFRP adherends should be prepared in the same manner as theplates to be bonded. The load and locus at failure is to be recorded andreported and the test should be deemed acceptable if the bond strength ofthe adhesive to the CFRP is greater than 12 Nmm�2.

It is recommended that representative samples of the prepared surface beused to make single lap shear joints generally in accordance with ASTM

SOR9 12/19/98, 4:04 PM261


D3163 (1992). The dimensions and arrangement of the test are shown inFig. 9.6. The failure load will indicate whether or not the adhesive materialhas been mixed and cured adequately, while the locus of the joint failurecan provide information on the soundness of the bond. Interfacial failureand low joint strengths indicate poor adhesion, whilst failure of the joint inthe composite material itself can at least provide some lower bound confi-dence level in bond performance.

An alternative to the single lap joint test is the wedge cleavage testillustrated in Fig. 9.7 and carried out generally in accordance with ASTMD3762 (1979). The purpose of the test is also to assess adhesion and tough-ness of the adhesive material. The procedure involves first bonding togethertwo strips of composite 150mm long and then, once the adhesive has cured,driving a wedge between them to provide a self-stressed specimen. Thisinduces a cleavage stress on the bondline and if adhesion is poor a crackwill grow between the composite and the adhesive layer. If the crack growsthrough the adhesive only, then satisfactory adhesion is assured; the frac-ture energy of the adhesive material can additionally be calculated from aknowledge of specimen geometry and material properties, enabling a checkon the toughness properties of the cured adhesive. The length of the crackand its position should be recorded.

Figure 9.6 Single lap joint (all dimensions in millimetres).

SOR9 12/19/98, 4:04 PM262


Figure 9.7 Wedge cleavage test (all dimensions in millimetres).

Tests should be carried out over a range of temperatures specificallyincluding �25 °C, �20 °C and �45 °C using CFRP adherends. The tempera-tures are measured by means of thermocouples attached to the CFRPsurface of the joint. The minimum average lap shear stress is 12Nmm�2 at20 °C.

9.6.2 Tests for new materials – epoxy resin adhesives forCFRP plate bonding

9.6.2.1 Materials

The material consists of a two-part cold cured, epoxy-based adhesive com-prising resin and hardener, the resin being based on the diglycidyl ether of‘bisphenol A’ or ‘bisphenol F’ or a blend of the two. The hardener or curingagent should be from the polyamine group in order to achieve a betterresistance to moisture penetration through the adhesive layer. Inert fillersmay also be incorporated with the resin component to improve the applica-tion or performance characteristics of the adhesive. The filler should be anelectrically non-conductive material, be highly moisture resistant, be able towithstand temperatures up to 120 °C without degradation and have a maxi-mum particle size of 0.1mm. The materials should be supplied in two packs,in correct weights for site mixing. The two components should have differ-ent colours in order to aid thorough mixing.

Other adhesives offering performance equivalent to that required by thisspecification for long term durability will be acceptable. In the absence ofsatisfactory laboratory accelerated durability tests, however, the assessmentof the durability of an alternative adhesive should be based on the provisionof evidence that it has, in practice, performed satisfactorily in conditionssimilar to the use proposed for a period of not less than 15 years afterinstallation.

SOR9 12/19/98, 4:04 PM263


9.6.2.2 Mixing

Mixing should take place in accordance with manufacturer’s instructions.

9.6.2.3 Placing

The adhesive should be capable of being applied readily to both concreteand CFRP surfaces in layers from 2–10mm thick.

9.6.2.4 Cure time and temperature

The adhesive should be capable of curing to the required strength between10 °C and 30°C in relative humidities of up to 95%. For repairs andstrengthening works the adhesive should cure sufficiently to give the speci-fied mechanical properties at 20 °C in not more than three days. On curingthe adhesive should undergo negligible shrinkage.

9.6.2.5 Usable life

Mixed adhesive, before application to the prepared surfaces should have ausable life in excess of 40min at 20 °C. Test method as in BS 5350 :Part B4(1993).

9.6.2.6 Open time

The time limit in which the joint should be made and closed should notexceed 20min at a temperature up to 20°C. Test method as in BS 2782 :Part8 Method 835 (1994).

9.6.2.7 Storage life

Shelf life of both the resin and the hardener should be not less than 6months in original containers at 5°C and at 25 °C.

9.6.2.8 Mechanical properties of hardened adhesive

Cure and storage temperature, as stated in the appropriate standards,should be used for all tests and a minimum of five tests should be under-taken for each test requirement and the average of the results taken.

9.6.2.9 Moisture resistance

Minimum moisture transport through the adhesive should be ensured, withwater content not exceeding 3% by weight after 28 days immersion in

SOR9 12/19/98, 4:04 PM264


distilled water at 20 °C. Specimens should be 40mm by 40 mm and between1–2 mm thick. The procedure and calculations are used in accordance withBS 6319 :Part 8, Clause 6.2.7 (1984) to determine water content.

9.6.2.10 Heat distortion temperature

The adhesive should have a heat distortion temperature (HDT) of at least40 °C. The sample under test should be placed in a temperature controlledcabinet at 20 °C and loaded in order to achieve a maximum fibre stress of1.81 MN m2 in accordance with ISO 75 (1987). The HDT should be takenas the temperature attained, measured by a thermocouple attached to thespecimen, after undergoing a further 0.25 mm deflection while subject to asurface heating rate of 0.5 °C min�1.

9.6.2.11 Flexural modulus

Flexural modulus, without creep effects, at 20 °C should be in the range of4–10 kN mm�2. A specimen 200 mm by 25 mm by 12mm deep tested in fourpoint bending should be used. The sample should be loaded at the thirdpoints to achieve a deflection at a rate of 1 mm min�1 and the centraldeflection recorded. From the resultant load–deflection curve, the secantmodulus at 0.2% strain is calculated.

9.6.2.12 Tensile strength

A minimum value of 12N mm�2 at 20 °C should be achieved. A dumb-bellspecimen should be manufactured in accordance with BS 2782 : Part 3(1996) with a cross-section of 10 mm by 3mm. The specimens should be castin polytetrafluoroethylene-lined moulds. Adhesive ductility may be meas-ured with appropriate strain monitoring equipment.

9.6.3 Concrete

The suitability of the preparation of the concrete surface should be demon-strated by a ‘pull-off’ test using the adhesive to be used in the bonding of theplates. The test should be deemed a failure unless the failure plane is withinthe concrete.

The bond of the plate to the concrete may be demonstrated after bondingby suitable tests such as coin tap, ultrasonic pulse velocity or thermography.

9.6.4 Trial panels and procedure trials

A trial panel and procedure trial for approval by the engineer should beprovided for the strengthening with a plate of the same width and thickness

SOR9 12/19/98, 4:04 PM265


and incorporating the same fixing details as those designed for the structure.This should demonstrate the following:

• preparation of the plate• preparation of the substrate surface• use of spacers• application of adhesive• installation of fixings• finishing of installation• compliance with quality control items

9.7 In-service inspection and maintenance

9.7.1 Manual of procedures

A manual of procedures for inspection and maintenance for the strength-ened structure should be prepared for the maintaining authority. A typicalsummary of inspection procedures is given below but this should not beregarded as exhaustive. A draft list prepared during design can be extendedduring construction should it become necessary.

It is suggested that inspections should take place every six months for atleast two years after completion of the work. The frequency of further in-spection should be agreed with the maintaining authority after a review ofthe inspection reports.

The inspection procedure manual should include all relevant technicalliterature relating to products used in the work. Photographs of criticaldetails and maps of cracking are also useful. Procedures to be adopted ifdeficiencies occur should also be clearly defined.

The deck soffit should be checked to ensure:

• There is no visible evidence of sealed cracks in the concrete opening.• There is no visible evidence of the plates debonding.• There is no audible evidence of the plates debonding by tapping the

plates with a coin or suitable light hammer.• The plate fixing bolts are not loose.

The top surface of the deck should be checked to ensure:

• There is no visible evidence of the road surfacing cracking or deforming.• There is no visible evidence of any cracks in the edge parapet detail

increasing in width or new cracks forming.

Hammer tapping must not damage the system or any nuts slackened onthe fixing system.

SOR9 12/19/98, 4:04 PM266


9.8 References

ASTM D3163 (1992) Standard Test Method for Determining Strength of AdhesivelyBonded Rigid Plastic Lap-Shear Joints in Shear by Tension Loading, ASTMPhiladelphia 15.06.

ASTM D3039/3039M (15.03) (1995) Test Method for Tensile Properties of PolymerMatrix Composite Materials, ASTM Philadelphia.

ASTM D3762 (1979) Standard test method for adhesive bonded surface durabilityof aluminum (Wedge test), ASTM, Philadelphia.

BS 1881 : Part 207 (1992) Recommendations for the Assessment of Concrete Strengthby Near-to Surface Tests, British Standards Institution, London.

BS 5400 : Part 4 (1990) Code of Practice for Design of Concrete Bridges – Steel,Concrete and Composite Bridges, British Standards Institution, London.

BS 5400 : Part 2 (1990) Specification for Loads, British Standards Institution,London.

BS 5350 : Part B4 (1993) Determination of Pot Life – Method of Test for Adhesives,British Standards Institution, London.

BS 2782 : Part 3 (1996) Mechanical Properties 1976, Method 320 B/C, TensileStrength, Elongation and Elastic-Modulus Method of Testing Plastics, BritishStandards Institution, London.

BS 2782 : Part 8 (1994) Other Properties 1980, Method 835C, Determination ofGelation Time of Polyester and Epoxide Resins using a Gel Time, BritishStandards Institution, London.

BS 6319 : Part 8 (1984) Method for the Assessment of Resistance to Liquids, BritishStandards Institution, London.

Chajes M J, Finch W W, Januszka T F and Thomson T A (1993), ‘Bond and forcetransfer of composite material plates bonded to concrete’, ACI J 93(2) 208–217.

CRAG (1988) Test Methods for the Measurement of the Engineering Properties ofFibre Reinforced Plastics, ed. P T Curtis.

EUROCOMP (1996) Structural Design of Polymer Composites – EUROCOMP –Design Code and Handbook, ed. J L Clarke, E & FN Spon, London.

Highways Agency (1994) Strengthening of Concrete Highway Structures Using Ex-ternally Bonded Plates, BA 30/94.

ISO 75 (1987) Plastic and Ebonite – Determination of Temperature of Deflectionunder Load.

Jansze W (1995) ‘Anchorage of externally bonded plates under static flexural load-ing’, Progr Concrete Res 4, February.

Jansze W and Walraven J C (1996) ‘Anchorage of externally bonded steel plates’,International Concrete Congress, In Service of Mankind, Dundee, Scotland, 24–28June 1996.

Jones R, Swamy R N and Charif A (1988) ‘Plate separation and anchorage ofreinforced concrete beams strengthened by epoxy-bonded steel plates’, The StructEng 66(5/1) 85–94.

Jones R, Swamy R N and Charif A (1989) ‘Plate separation and anchorage ofreinforced concrete beams strengthened by epoxy-bonded steel plates’, The StructEng 67(13/4) 250.

Kaiser H (1989) Strengthening of Reinforced Concrete with Epoxy-Bonded Carbon-Fiber Plastics, Thesis submitted to ETH, Zurich, Switzerland, in partial fulfilmentof the requirements for the degree of Doctor of Philosophy (in German).

SOR9 12/19/98, 4:04 PM267


Meier U (1987) ‘Bridge repair with high performance composite materials’, J MaterTechnik 4 125–128.

Meier U and Kaiser H (1991) ‘Strengthening of structures with CFRP laminates’,Advanced Composites Materials in Civil Engineering Structures Proceedings, MTDiv/ASCE/Las Vegas, Jan. 31 1991.

Neubauer U and Rostasy F S (1997) ‘Design aspects of concrete structures strength-ened with externally bonded CFRP-plates’, Structural Faults & Repair – 97,Edinburgh, July 1997, Vol 2, pp 109–118.

Oehlers D J and Ahmed M (1996) ‘Upgrading reinforced concrete beams by boltingsteel side plates’, Bridge Management 3, Inspection, Maintenance and Repair,Guildford, 1996, E&FN Spon, pp 412–419.

Poulsen E (1996) Anchorage of Sika“ CarboDur Laminates Bonded to Concrete,CEN TC104 SC8 WG3 meeting in Paris 5/6 Dec 1996.

prEN 1542 Products and Systems for the Protection and Repair of Concrete Struc-tures – Test Methods, the Pull-off Test, Awaiting German translation beforeratification, not available publicly.

Ranish (1982) Zur Tragfähigkeit von Verklebungen zwischen Baustahl uns Beton –geklebte Bewehrung, Institute for Construction Materials, Massive type Construc-tion and Fire Protection, Technical University Braunschweig, Issue 54.

Raoof M and Zhang S (1996) ‘Peeling failure of reinforced concrete beams withexternally bonded steel plates’, Bridge Management 3, Inspection, Maintenanceand Repair, Guildford, 1996, E&FN Spon, pp 420–428.

Regan P (1993) ‘Research on shear: a benefit to humanity or a waste of time?’ TheStruct Eng 71(19) 337–346.

Roberts T M (1989) ‘Approximate analysis of shear and normal stress concentra-tions in the adhesive layer of plated RC beams’, The Struct Eng 67(12/20) 229–233.

Rostasy F S (1993) ‘Strengthening of R/C and P/C structures with bonded steel andFRP plates’, Proc. 5th Int. Conf. On Structural Faults and Repair 1993, Vol 2, pp217–224.

Swamy R N, Jones R and Charif A (1989) ‘The effect of external plate reinforce-ment on the strengthening of structurally damaged RC beams’, The Struct Eng67(3/7) 45–56.

Täljsten B (1994) Strengthening of Existing Concrete Structures with Epoxy BondedPlates of Steel or Fibre Reinforced Plastics, Doctoral Thesis, Luleå University ofTechnology, Division of Structural Engineering, Sweden. August 1994.

Tausky (1993) Betontragwerke mit Aussenbewehrung, Birkhäuser Verlag.Taylor H P J (1974) ‘The fundamental behaviour of reinforced concrete beams in

bending and shear’, Proc. ACI-ASCE Shear Symposium, Ottawa, 1973 (ACIPublication SP42-3) American Concrete Institute, Detroit, 1974, pp 43–77.

Triantafillou T C and Plevris N (1992) ‘Strengthening of RC beams with epoxy-bonded fibre-composite materials’, Mater Struct 25 201–211.

Wu Y S and Gibson A G (1994) ‘Composites offshore: applications and opportuni-ties’, Composites 94, Birmingham 22/23 November 1994, British Plastics Federa-tion Publication No 293/10 pp 45–56.

Zhang S, Raoof M & Wood L A (1995) ‘Prediction of peeling failure of reinforcedconcrete beams with externally bonded steel plates’, Proc Inst Civ Eng StructBuilding 110(Aug) 257–268.

SOR9 12/19/98, 4:04 PM268


9.9 Bibliography

BD2 (DMRB 1.1) Technical Approval of Highway Structures on Motorways andOther Trunk Roads.

BA 35 (DMRB 3.3) The Inspection and Repair of Concrete Highway Structures.SB 1 (DMRB) (Scotland) The Inspection of Highway Structures.DB 24 (DMRB 3.1) Design of Concrete Highway Bridges and Structures, Use of

BS5400 : Part 4: 1990.DB 37 (DMRB 1.3) Loads for Highway Bridges.BD 21 (DMRB 3.4) The Assessment of Highway Bridges and Structures.

SOR9 12/19/98, 4:04 PM269


270

10Site construction techniques

A P R I M O L D I

10.1 Introduction

The bonding of plates to civil engineering structures has been recognisedas useful since the end of the 1970s. The work is usually carried out in orderto enhance the load-bearing capacity of the structure, either because thedesigned loading is likely to be exceeded or because modification ordeterioration of the structure demands it. This chapter describes the tech-niques and problems associated with plate bonding in real commercialsituations on construction sites.

Concrete Repairs Ltd is a well established specialist contracting companywith a wealth of experience in the repair, maintenance and modification ofreinforced concrete structures. It has been carrying out plate bonding sincethe early days of acceptance of the technique in the UK marketplace. Since1996 the company has carried out several projects using carbon fibre rein-forced polymer (CFRP) plates as a strengthening medium for reinforcedconcrete structures.

Experience gained in the technique of externally bonding CFRP plates toconcrete structures, by undertaking commercial contracts over a two yearperiod, is summarised in the following sections. This technique is the oneadopted by the ROBUST system.

10.2 Steel plate bonding

Before discussing the use of CFRP plates it is important to consider brieflythe more traditional technique of externally bonding steel plates.

Since the strengthening in 1975 of the Quinton Bridge on the M5motorway, steel plate bonding has been increasingly used in the UK forupgrading concrete beams. The concrete in which steel is embedded asreinforcement can serve to protect the steel because the highly alkaline

SOR10 12/19/98, 4:21 PM270

Site construction techniques 271

nature of concrete provides an environment that is beneficial to the embed-ded reinforcement. If the concrete is of good quality, well compacted andwith minimal cracking, it is difficult for the process of corrosion to takeplace. In such circumstances, reinforced concrete can support the tensileand shear stresses experienced by the structure.

However, corrosion of steel reinforcement through the process of car-bonation can be one of the major limitations to the longevity of a reinforcedconcrete (RC) structure, with the presence of chlorides, poor detailing andworkmanship all contributing to the corrosion process. In extreme casesof degradation a structure may become unsafe or beyond economic repair.There is a continuing worldwide problem regarding corrosion of embeddedreinforcement, and of course, a worldwide industry maintaining structuresand preventing corrosion.

In most cases, externally bonded reinforcement (in the current contextbonded plates) is required to add to the capabilities of the existing embed-ded reinforcement. This can be to enhance strength or effectively to replacecorroded reinforcement.

In the past, upgrading structures was achieved by literally adding a fur-ther layer of reinforcement to that of the existing steel and then encasingthe whole within a concrete jacket. This was a very expensive process,tantamount to rebuilding a large section of the structure.

The idea of adhesive bonding steel plate to a suitably preparedconcrete surface was then developed. This appeared to be a relatively easymethod for construction sites, but in practice handling up to 6m long plates,usually on to soffits of beams, and in addition providing multiple endanchors without damaging the applied coating on the plates can be anextremely difficult operation. The survey and setting out of the end anchorbolts is fraught with difficulty especially on a heavily reinforced structure.The protective paint coatings to the plates need to be applied under almostclinical conditions and, of course, the coatings require maintenance duringservice life. Lack of maintenance of the plates will result in their rapidcorrosion as they will not then have the benefit of being embedded inconcrete.

Steel plate bonding is a perfectly viable means of providing extrastrength. However, the logistics of handling plates weighing severalhundred kilograms makes the operation a difficult one. Furthermore, fabri-cation and preparation processes need to be allowed for when program-ming works and the slow application often means lengthy restrictions in theuse of the structure. This could result in long traffic delays when dealingwith road bridges.

The use of CFRP plates simplifies and shortens the whole process ofplate bonding and its adaptability allows a wide range of options to beconsidered.

SOR10 12/19/98, 4:21 PM271


10.3 Adhesive bonding of carbon fibre composite

plates – site requirements

The following sections will discuss the management, design and site con-struction details which are associated with adhesively bonding CFRP platesto RC beams.

10.3.1 Construction Design Management Regulations1994 (CDM)

Virtually all commercial construction projects in the UK need to be carriedout under CDM (1994) regulations (this requirement became active fromMarch 1995). These specific legally enforceable regulations need to beconsidered and followed at every stage in the design and constructionprocess. The regulations concern the management of health and safety. It isassumed that all works described in this chapter properly comply with theRegulations.

10.3.2 Health and safety

It is imperative that all construction site operations are carried out in sucha manner that risk to site workers and others that may be affected by thework is either eliminated or reduced to a safe level. It is assumed thereforethat in all the following descriptions risk assessments have been madeand that where necessary personal protective equipment (PPE) is suppliedand used. It is also assumed that a safe means of access and materialshandling is employed and that suitable measures to protect passers-by aremaintained.

10.3.3 Site set-up

Before carrying out any other works the site should be visited to determineits location, its relationship with the immediate environment, ease of accessfor personnel and deliveries, provision of services, security and other spe-cial considerations to enable a work site to be established.

Following this, a suitable works compound should be built provid-ing accommodation and facilities for personnel, secure weatherproof storagefor materials and, of most importance, a clear unobstructed area for layingout the plates. Consideration must be given to the prevailing weatherconditions, time of year and any other physical factors that may affect theworks.

Finally a safe means of access to the working area if necessary, must be

SOR10 12/19/98, 4:21 PM272


built. This must take due account of risk of falling, work over water or liverailway tracks and must be constructed accordingly.

10.3.4 Loading

The designer of the strengthening system may need to take due account ofdead and imposed loadings during the bonding process. For example, ahighly trafficked road or rail bridge may introduce stresses into the platesduring the curing process. If the designer finds that this is unacceptable,several other options are available. Traffic can be removed from the struc-ture either by lane restrictions, night-time possession or total closure. Alter-natively, propping may be installed to accept the loading during thebonding process. Other forms of preload, such as jacking or kentledge, maybe applied as a method of inducing a stress into the plates as part of thedesign and installation process.

10.3.5 Condition survey

Once the structure can be easily inspected a comprehensive concrete condi-tion survey should be carried out. This starts with a visual inspection tocheck for obvious defects (cracks, spalls, poor compaction, uneven sur-faces) and is followed by a hammer sounding survey to identify hollowareas. Additionally, chemical tests (depth of carbonation, chloride levels)may be necessary.

Prior to carrying out the bonding works it is important that a sound struc-ture is available. The previously mentioned tests will identify defects anddeterioration. If necessary a package of repair can be carried out to restorethe structural concrete to its original state and thereby create an undamagedstructure for the bonding works. If anchor bolts are required, an electroniccovermeter can be used to identify the position, size and depth of embeddedreinforcement. All the information gained from the condition survey shouldbe carefully drawn up and recorded for future reference and use.

10.3.6 Setting out

Should the design require it or if there are special considerations (e.g.prestressing) then setting out should be considered at this stage. This willenable the anchor plates to be bonded into position and the positions of theanchor holes on relatively flat surfaces to be marked. It may be necessary toclean the surfaces by high pressure water jetting to remove dirt and con-taminants to make the marking up process easier. The setting out is onlyrequired at this stage if hole drilling and coring is to be carried out beforeplates are bonded.

SOR10 12/19/98, 4:21 PM273


10.3.7 Hole drilling

Where predrilled or prestressed plates are to be used it is advisable toconsider hole drilling at this stage. Using the information from thecovermeter survey and suitable templates, the positions for end anchoragebolts or prestressing fixings can be marked accurately. If the design allowsthe existing reinforcement to be cut, then slow speed diamond coring can beused. This process uses water as a lubricant and will provide an accuratecore hole. Any intervening reinforcement will be cut in the process.

Percussion drilling is usually quite satisfactory, provided that accuracycan be maintained. Intervening reinforcement will not be removed by thisprocess, so careful positioning and planning is required. This is especially sowhere a group of bolts (often 6 or 9 no.) is required in a small area. Theperpendicularity of the holes is also extremely important. Although dia-mond drilling is usually carried out using a jig, percussion drilling is oftenhand held. Slight deviations or inaccuracies at this stage will mean that itbecomes difficult or impossible to position the anchor bolts through theplates. Particular care must be taken when drilling into prestressed or post-tensioned structures to avoid damage to the tendons.


To ensure successful bonding of the plates, a high standard of surfacepreparation is essential. Concrete surfaces should be free of all contami-nants including grease, oil and existing coatings. Surface laitance must alsobe removed, as should any unevenness or sharp ridges and formwork markson the concrete surface. Sudden changes in level should be less than 0.5mm.

A well-weathered concrete surface could be suitably prepared using highpressure water jetting techniques. All other surfaces (which, if recentlyrenovated should be at least 6 weeks old) should be prepared by grit-blasting followed by vacuum cleaning (or vacuum blasting if available).

If gritblasting is not possible, scabbling, needle gunning or grinding maybe utilised, although these techniques are not recommended as they willalmost certainly leave a much rougher surface than gritblasting. Such sur-faces may require remedial works before plate bonding can begin. Signifi-cant surface imperfections can therefore be dealt with at this stage. Usuallyan epoxy mortar would be used, which provides a fast gain in strength andwill allow overbonding to take place quickly.

As a final check on the surface preparation, it is usual to perform aminimum of three pull-off tests on the prepared substrate; these tests willalso provide an indication of the concrete strength. The test method in-volves bonding a 50mm diameter steel dolly to the concrete surface afterpreparation.

SOR10 12/19/98, 4:21 PM274


The load at pull-off and the location of the failure plane should berecorded. This simple test will help to ensure good surface preparation andto give an indication of the tensile strength of the concrete substrate.

10.3.9 Working area

At this point in the proceedings it is worth considering the working area.The CFRP plates are extremely light and often very long (20m). Whenhandled they will be coated with adhesive which will pick up any contami-nants that may come into contact with them. A clear unobstructed workingarea is therefore essential. Consideration should be given to adverseweather conditions that may affect the work (e.g. wind, rain, low tempera-tures). If necessary, temporary roofing or heating should be provided. Itmust be possible to transport the coated plate to the structure without anyrisk of it becoming contaminated.

A suitable source of electricity should be available together with cleaningand washing facilities for the plant and personnel. Lighting should also beavailable if necessary.

10.3.10 Materials and plant

CFRP plates come in varying widths (e.g. 50 mm, 90mm, 100mm), varyingthickness (1.0mm, 1.2 mm) and any required length. CFRP plate is veryexpensive and therefore minimisation of waste is an important factor. Rolllengths of 250 m are available. Wastage from rolls must be considered whenordering cut lengths. Plate can be easily cut to length on site using aguillotine.

Some CFRP plate is delivered in its raw state and will require cleaningto remove contaminants. Other types of plate are supplied with a peel-ply protective coating which is removed immediately prior to bondingto ensure a completely clean surface. Plate should be checked fortwist. A twisted plate will not lay flat on the surface and should berejected.

CFRP plate should be handled carefully. The unidirectional longitudinalweave of the fibres means that it can easily split along its length. Thischaracteristic is far less of a problem on peel-ply plates.

Prestressed plates may well be supplied with glass fibre, predrilled anchorblocks already bonded to cut lengths of CFRP. Again, the dimensions andquality of these must be checked before use as an incorrectly supplied platecan cause delay to the contract and be costly to remedy.

Epoxy adhesives are available from several sources. They are two part(resin and hardener) for site mixing. Again material choice is based onseveral factors. Client or contractor preference, track record, technical

SOR10 12/19/98, 4:21 PM275


performance and cost. Special considerations such as pot life, ambienttemperature and environmental issues may also influence the choice.Special care must be taken when disposing of the used containers.

It is usual, during the course of a project, to take samples from each batchof the adhesive used. These specimens are usually 200mm long by 12 mmdeep by 25mm wide and formed in steel moulds. They should be cured inthe same ambient conditions and for the same length of time as that usedfor the bonding operation. After demoulding, these samples are tested anda load/deflection graph is recorded. Tensile strengths of adhesive can bechecked after manufacturing a dumb-bell specimen which is loaded tofailure. These tests are further detailed in BA 30/94 Part 1 (1994). Allmaterial testing should be carried out by an independent testing authority(Cranfield University, 1996). The ROBUST project recommends that lapshear tests should be undertaken (see Chapter 3, Section 3.3.6).

Ancillary equipment such as anchor bolts, sleeves and studs are usuallychosen from manufacturers’ catalogues after considerations of exposure(stainless steel, plated steel) and technical specifications.

Plant is generally confined to fairly small items. Slow speed mixing drillswith paddles for the adhesive are usually employed together with handtoolsfor adhesive application. Drilling machines will be necessary for anchor boltholes if they are to be formed after the plates are bonded.

As in all parts of the project the correct personal protective equipmentmust be available and used at all times.

10.3.11 Bonding of unanchored plates

The simplest form of CFRP bonding is a single layer unanchored plate.These are installed when the designer considers that end anchorages areunnecessary.

The plate is selected and laid out upside down on the adhesive table. If ithas a peel-ply protective coating this is removed from the adhesive side atthis stage. If it is an unprotected plate, its surface is lightly sanded andcleaned using a solvent wipe. The plate is visually inspected for defects andimperfections. From this point on cleanliness is vital and operatives wearsurgical-type gloves. The concrete surface is finally checked and lines aremarked to show the approximate position of the plate.

Adhesive mixing can now take place. The cold cured epoxies usually haveresin and hardener in differing colours (e.g. white and black). When mixedtogether with a slow speed paddle drill the mixture turns into a grey paste(Sika CarboDur, 1998). Mixing is complete when there are no inconsisten-cies in colour.

Mixing usually takes place in the larger of the two delivery pails. The slowspeed mixing paddle avoids air entrainment. Depending on ambient condi-

SOR10 12/19/98, 4:21 PM276


tions the pot life is about 30 min. Before applying adhesive a final check ofall surfaces and components should be made.

The mixed adhesive is applied by hand trowel to the selected position onthe concrete structure (Fig. 10.1). Using skilled labour, conventional plas-tering and scraping techniques will ensure that all voids are filled and a1 mm thick skim adhesive is left on the surface.

During this operation it is recommended the mixed adhesive is alsoapplied to the plate. It is suggested that a ‘dome’-shaped profile of adhesiveis formed, some 2mm thick in the centre of the plate, thinning to 0.5mm atthe edge. This can be achieved using a dome-shaped spatula fixed at oneend of the adhesive reservoir. The whole assembly can be drawn along thelength of the plate, thus extruding the required profile of adhesive to theplate. This is a quick and easy method (Fig. 10.2).

Spacers may be required to maintain adhesive thickness, although theseare not essential. Conventional plastic washers of known thickness can beset into the adhesive. Alternatively, single sized glass balls or ballotini canbe sprinkled over and embedded into the adhesive. Both systems are suc-cessful at maintaining a minimum adhesive thickness.

The plate is then ready to be applied to the concrete surface. This opera-tion usually requires three operatives if the plates are more than 3m long.The coated plate is lifted from the table and taken to the structure. One endis placed on the surface using only hand pressure. Working away from thatend and using hand pressure only, the plate is bonded to the structure.

Figure 10.1 Applying epoxy adhesive to prepared concrete surface.

SOR10 12/19/98, 4:21 PM277


Figure 10.2 Setting dome-shaped spatula into adhesive reservoir priorto application of epoxy adhesive to plate.

Moving along the plate the technique continues until the far end is reached.Even when working on soffits the adhesive is strong enough to take thevery light weight of the CFRP plate without risk of sagging or debonding(Fig. 10.3).

Once in place a hard rubber roller is applied to the plate to squeeze outexcess adhesive and ensure firm adherence to the substrate. A trowel isused to tidy up the edges and remove excess adhesive. This adhesive shouldnot be reused (Fig. 10.4).

The bonding of one plate is thus completed. The exercise, after thoroughplanning, is very straightforward and can then be repeated for multiplateapplications.

After 48h the fitted plate should be checked for voids. One very simpleand effective method is to ‘coin-tap’ the surface of the composite plate.More sophisticated methods would be the thermography, ‘woodpecker’

SOR10 12/19/98, 4:21 PM278


Figure 10.3 Fitting CFRP plate to concrete surface.

Figure 10.4 Using a hard rubber roller to squeeze out excess adhesiveand ensure firm adhesion.

and ultrasonics techniques. If voids are found, these can be injected withepoxy resin through the sides of the exposed adhesive. This method issatisfactory for steel plates but is not recommended for CFRP plates. Itmust be stressed, however, that voids in the adhesive layer are extremelyrare. If peel-ply is fitted to both sides of the composite plate the outer peel-

SOR10 12/19/98, 4:21 PM279


ply layer can be removed after the composite plate has fully bonded to theadherend.

10.3.12 Bonding of anchored plates (predrilled)

There are circumstances when, for design considerations, the CFRP platesare supplied with anchor bolt plates in fibreglass already bonded to theends and predrilled. As before the chosen plate must be offered up tothe predrilled concrete to check dimensions and hole positions. Once ap-proved, the anchors for one end only can be fitted. These may be singleanchors or a group.

Again, depending on design and anchor characteristics, expanding orresin anchors can be used. The former expand and grip the predrilledconcrete using expanding wedges. The latter use epoxy or polyester resin inpremixed or capsule form to set, either a threaded sleeve or a stud into theconcrete.

The setting of these anchors is critical. They are often of at least 12mmdiameter and sometimes 25mm or 30mm. They must be set centrally intheir holes and perpendicular to the concrete surface. Extra care must betaken where multiple anchors are fitted to ensure that they are correctlyspaced in relation to one another (Fig. 10.5).

Figure 10.5 Detail showing bolted anchorage of a multiple plateinstallation.

SOR10 12/19/98, 4:21 PM280


Once the anchors are set for one end of the plate the bonding of the platecan take place. The techniques are similar to those for unanchored plates.The coated plate is placed onto its anchor bolts and the nuts and washersare hand tightened. Using hand pressure only, and working away from theanchored end, the plate is bonded to the structure. When the holes in theplate and structure are aligned, the anchor bolts at the other end can beinstalled and again be hand tightened. Once in place pressure is applied tothe plate using a hard rubber roller to squeeze out the excess adhesive andensure firm adhesion to the substrate. After 48h the plate can be checkedfor soundness and voids and the anchor bolts can be tightened to therequired torque.

10.3.13 Bonding of anchored plates (undrilled)

If anchor systems are required the most frequently used technique duringthe installation of CFRP plates is to apply end anchor bolts to undrilledplates. (ROBUST does not recommend this method as it can weaken thecomposite plate. Section 4.3.3 of Chapter 4 discussed the ROBUST inves-tigative work on anchor plates.)

In this procedure the plate length and the setting out of the anchoragesystem is not nearly so critical as when using a predrilled plate. The tech-nique for bonding is identical to that of the unanchored plates and is,therefore, much quicker than a predrilled plate system. Any site tolerancescan also be accommodated. Once the plate has been successfully bonded tothe structure it can be drilled to accept its anchors. Care must be taken notto hit embedded reinforcement. A covermeter should be used to ascertainthe position of the embedded reinforcement. The CFRP plate is ‘invisible’to a covermeter thus making the setting out relatively straightforward.

The CFRP plates are easily drilled in situ using a hand-held rotary drillset at slow speed. Again, the setting of the anchors, checking and testing isthe same as for the previous sections.

10.3.14 Bonding of multiple plate thicknesses, anchoredand unanchored

There are projects that will require a multithickness plate installation toprovide the required increases in structural strength. Whether these instal-lations are anchored or unanchored, the techniques are similar to singlethickness work.

After bonding the first layer to the concrete structure the installedplate becomes the substrate for the next layer, and so on. The peel-ply tothe exposed surface of the composite plate is removed. If a plate withoutpeel-ply has been fitted then the surface of that plate should be lightly

SOR10 12/19/98, 4:21 PM281


Figure 10.6 Detail showing spliced plate application.

sanded to provide a key. The same resin adhesive is used and the next plateis bonded.

If anchors are necessary they can be installed through predrilled holes orholes can be drilled through the multilayers of plates at the end of eachgroup of plates; however, ROBUST does not recommend this procedure.The technique can also be used to strengthen a structure in two or moredirections by criss-crossing the plates (Fig. 10.6).

If drilling the main plates is not possible then short lengths of CFRP platecan be used to provide an end anchorage as a strap perpendicular to themain plate.

10.3.15 Prestressing of CFRP plates

The final option available to both the designers and installers of CFRPplates is prestressing. This is a new technique that has only been carried outonce on a full size structure. However, the success of the process will allowthe use of CFRP plate strengthening techniques to be considered for awhole range of projects.

The prestressing machine is essentially a custom built plate frameworkthat is securely bolted to the concrete substrate. Attached to the frame aretwo manual screw jacks which in turn are connected to a roller mountedstressing bar. This bar is locked to a stressing block at one end of the CFRPplate using a shear pin (Fig. 10.7).

SOR10 12/19/98, 4:21 PM282


Figure 10.7 Prestressing machine bolted in place to a reinforcedconcrete beam.

Figure 10.8 Connecting the CFRP plate to the prestressing machine.

A predetermined axial stress is applied to the plate using simple load/extension considerations, that is, a known extension will correspond to acalculated load. Adaptations of the machine are possible for clamping tobeams or soffits and for stressing two plates simultaneously (Fig. 10.8).

SOR10 12/19/98, 4:21 PM283


Figure 10.9 Using the twin screw jacks to tension the plate and applyprestressing.

The prestressing machine has to be accurately and securely fixed to thestructure. The anchor bolts for the machine have to be carefully set out andfitted, not only to facilitate secure mounting of the machine under load butto maintain the correct relationship with the anchor bolt that will secure thestressed plate.

After calculating the elongation of the plate under load, the end anchorposition after stressing can be calculated and set out. The end anchor boltscan be set into the concrete substrate and the adhesive applied to the plateand substrate as before. The anchor block on the plate is then connected tothe shear pin on the prestressing machine.

After checking that all the connections are secure the stressing processcan begin. The slack is taken up using the screw jacks and other adjust-ments. The starting position of the taut plate in relation to a fixed scale isrecorded. Using the two screw jacks the plate is carefully tensioned (Fig.10.9).

The load required in the plate will be reached at a calculated extension.This is measured on a fixed scale. When the required extension (and there-fore load) is reached the anchor bolt holes on the plate will correspond tothose predrilled on the structure. When they coincide the anchor bolts canbe inserted and tightened (Fig. 10.10). The screw jacks can then be released,the stressed plate being securely bolted to the structure. Bearer beams andprops may need to be fitted against the plates at this stage to ensure firmadherence to the substrate (Fig. 10.11).

SOR10 12/19/98, 4:22 PM284


Figure 10.10 Inserting the anchor bolts after applying the requiredprestress load.

Figure 10.11 Props and bearer beams installed to prestressed plate.

SOR10 12/19/98, 4:22 PM285


Finishing and testing of the installed prestressed plate will be as before.Again, multiple thickness installations are possible with this technique.

10.3.16 Maintenance and protection

Once installed the CFRP plates require no maintenance. There is obviouslyno risk of corrosion of the plate itself. The use of stainless steel anchor boltswill ensure a very long life in even the harshest of environments. There arefew circumstances that would prevent the design life of the strengtheningsystem equalling that of the structure.

After removal of the access required for the installation, the plates areusually inaccessible. However, there are circumstances where the platesthemselves may need protection or hiding for purely aesthetic reasons. Itmay be necessary to provide fire protection to the installation or to cover itup to avoid the risk of vandalism.

The installed plates can easily be covered with a cementitious overlayapplied by spray or hand placed. To ensure adhesion of the overlay, anepoxy bonding coat is applied prior to the cementitious material. Theoverlay can be left ‘as sprayed’ or trowel finished smooth and, if required,subsequently painted. The strengthening system can therefore be hiddenfrom view very successfully and with great confidence that future mainte-nance inspections are unnecessary.

It is important that the installation is noted in the site safety file (arequirement of CDM (1994)), especially if it is covered up. This will pre-vent unauthorised drilling or damage to the plates from subsequent worksand trades.

10.4 Economics

In comparison to the steel alternative, CFRP plates are certainly not cheap.The raw material cost is often four times that of steel. However, the advan-tages are found in installation costs and whole life costings.

The costs of installation in terms of labour are much lower than for steel.Transport and handling costs are lower and, of most importance, the instal-lation of CFRP plate is very fast. A reduced contract programme obviouslylowers the ancillary costs of access and plant hire, propping and site set-up.Even more significant are the reduced timescales for road closures or trafficmanagement.

With negligible planned expenditure on maintenance, the economicsof CFRP plates become very attractive. Currently, the high cost of thecarbon fibre composite is offset by the advantages detailed above. Thisresults in CFRP plates being a viable and realistic alternative to the utilisa-tion of steel.

SOR10 12/19/98, 4:22 PM286


Figure 10.12 Overall view of completed installation of plates to areinforced concrete bridge beam.

10.5 Conclusion

Structural strengthening using externally bonded plates is a well-established process in construction and repair. The acceptance of CFRPplate as an alternative to steel has opened up the marketplace (Fig. 10.12).It is a straightforward and elegant technique whose performance can beeasily monitored and controlled in a site environment. The flexibility of thesystem allows the designer a greater scope to achieve his goals. The simplic-ity of the system allows the installer to work in a risk free environmentwithout the problems associated with handling and maintaining heavy steelplate. The economics of the system now compare favourably with those ofsteel and must surely be attractive to the client when the need for futuremaintenance is removed.

10.6 References

BA30/94 (1994) Design Manual for Roads and Bridges, Vol 3, Section 3, Part 1,BA30/94, Strengthening of Concrete Highway Structures Using ExternallyBonded Plates. Department of Transport, Feb 1994.

CDM (1994) Construction (Design & Management) Regulations (CDM), SI 1994/3247. HMSO.

Cranfield University (1996) Quality Control Testing Associated with Plate BondingWorks, Internal Report, Royal Military College of Science, Shrivenham. January1996.

Sika CarboDur (1998) Data Sheet 1.95, Sika, Welwyn Garden City, Herts, UK.

SOR10 12/19/98, 4:22 PM287


288

11Case studies of carbon fibre bonding

worldwide

M A SHAW AND J F DREWETT

11.1 Introduction

The following case studies utilising the carbon fibre/polymer compositestrengthening concept have been incorporated into this book to illustratesome practical examples of different strengthening applications. TheROBUST project which has recently been completed has not, as yet,any proven practical examples of the strengthening of existing bridge orbuilding structures.

Sika Ltd, Hertfordshire, UK, one of the members of the ROBUST Con-sortium, has developed over a number of years a procedure for strengthen-ing, using a system approach comprising an advanced epoxy structuraladhesive and a range of carbon fibre/epoxy polymer laminates. This com-bined system is known as the Sika CarboDur system.

Sika have been in the forefront of structural adhesive engineering in theconstruction industry for many years. Structural adhesives have been spe-cifically developed to transfer stresses from one substrate to another for arange of different situations. Typical applications include segmental bridgeconstruction and strengthening structures using external plate bondingtechniques. Each material has undergone intensive research and develop-ment and site trials to comply with the industry’s demands of cost effective-ness, performance and durability.

In the past external strengthening had been successfully carried out withsteel plates and epoxy-based structural adhesives. There are limitations toboth the design and practical aspects when using plates and these have beenhighlighted in the earlier chapters. Sika have been involved with over 200CarboDur projects worldwide, many of which would not have been possibleusing traditional strengthening techniques. The case histories selectedshow the adaptability and cost effectiveness of the technique and the globalacceptance of the CarboDur system.

SOR11 12/19/98, 4:24 PM288

Case studies of carbon fibre bonding worldwide 289

11.2 System properties

The general properties of the system are shown below. The carbon fibrelaminates are manufactured by the pultrusion process to the Swiss FederalLaboratories for Material Testing and Research (EMPA) specification.

11.2.1 Sika CarboDur laminates

Sika CarboDur laminates consist of:

• carbon fibre reinforced with epoxy matrix• fibre volumetric content �68%.

The laminates are available in a range of different grades. The minimumand maximum properties are shown in Table 11.1. Intermediate grades liewithin this range.

11.2.2 Structural adhesive, Sikadur 30

The properties of Sikadur 30 are

• two-part solvent free epoxy adhesive• density 1800kg/m�3

• glass transition point about 62 °C• flexural modulus 12 800 Nmm�2

• compressive strength about 90Nmm�2

• tensile strength about 30Nmm�2

• tensile slant shear strength about 18Nmm�2

• moisture uptake �0.5%.

For the case histories included in this chapter, the application of theCarboDur system has generally followed the procedure outlined below:

• The substrate surfaces were prepared to remove all contaminants.– Concrete: fine and coarse aggregate exposed and fine gripping

texture achieved by blast cleaning,

Table 11.1 Minimum and maximum properties of Sika CarboDur laminates

Minimum Maximumproperties properties

Thickness (mm) 1.2 1.4Width (mm) 50 150E-modulus (Nmm�2) 165000 300000Ultimate tensile strength (Nmm�2) 1450 3050

SOR11 12/19/98, 4:24 PM289


– Timber: surfaces planed or grit blasted,– Masonry: fine gripping texture achieved by blast cleaning or

scabbling.• Substrate surfaces vacuum cleaned to remove dust.• CarboDur carbon fibre laminates cut to length at factory or on site.• Preprepared laminate bonding surface cleaned.• Sikadur 30 structural adhesive applied to laminate and substrate.• Carbon fibre laminate pressed into position on substrate by hand.• Carbon fibre laminate bedded into adhesive using hard hand rubber

roller to extrude excess adhesive and produce void free bondline.• Surplus adhesive removed from substrate and laminate.

11.3 Case histories

11.3.1 Kings College Hospital, London, UK

11.3.1.1 The problem

In June 1996, this was the first project in the UK to use a carbon fibrestrengthening system for external poststrengthening. The project involvedthe overall refurbishment and extension of the Normanby College suitefor the new Joint Education Centre. As part of the refurbishment brief, anextra floor was required for accommodation. This was achieved by convert-ing the roof slab to a floor slab thereby increasing the live load capacity to3.0kN m�2. The construction of the existing building consisted of a rein-forced concrete frame with cast in situ troughed floors. The upper storeyextension was designed as a lightweight steel frame structure, howevercalculations on the existing roof slab established that in its present form, theslab could not sustain the additional live load.

To resolve the problem, three options were considered:

• Demolish the existing roof and construct a new floor.• Provide a secondary steel frame to support existing slab or new separate

floor.• Externally poststrengthen the roof slab by external plate bonding.

The first option was considered time consuming and expensive. Thesecond option was not feasible owing to the long spans. Thereforethe option of poststrengthening was preferred, based on its cost effective-ness and speed of application. The initial design showed that steel plates75mm wide and 6.0 mm thick were required to provide the additionalreinforcement.

SOR11 12/19/98, 4:24 PM290


11.3.1.2 The solution

The original roof slab was formed from 400 mm deep tapered ribs 80mmwide at the bottom located at 600 mm centres spanning 11.0 m. Because theplate width to thickness ratio was less than 50 and it would be necessary toslice the plates for handling purposes within a restricted working area, theuse of the CarboDur strengthening system was specified by the structuralengineers, Lawrence Hewitt Partnership. To achieve the desired strength-ening requirements, the laminates used were 50 mm wide, 1.2mm thick withan E-modulus of 155 000 Nmm�2.

In competitive tender, Concrete Repairs Ltd (CRL) secured the subcon-tract to supply and install the CarboDur system for the main contractorJohn Mowlem Construction Plc.

CRL prepared the floor rib soffits by needle gunning and vacuuming;this method of preparation was checked on site by performing pull-offtests using a ‘limpet’. The pull-off values achieved were in the region of3.0 N mm�2 with a failure in the concrete. The 50 mm wide CarboDurlaminates arrived on site in 250 m long rolls, preboxed for protection (Fig.11.1). The individual laminates were then cut on site using a guillotine,cleaned to remove surface contaminants and coated with the epoxy adhesive.

At the same time the concrete bond surface was coated with adhesive(Fig. 11.2). The laminates were then offered up to the beams (Fig. 11.3) and

Figure 11.1 The complete CarboDur strengthening system.

SOR11 12/19/98, 4:24 PM291


Figure 11.2 Application of adhesive to CarboDur laminate.

Figure 11.3 Positioning an 11.0 m laminate to underside of rib.

SOR11 12/19/98, 4:25 PM292


Figure 11.4 Bedding CarboDur laminate into adhesive with hand roller.

rollered into position (Fig. 11.4) to ensure good adhesive contact and toeliminate voids. The next day antipeel bolts were installed at the ends of thelaminate by drilling through the CarboDur laminate into the concrete andusing chemical anchors. The strengthening work was completed within thefour weeks programme at a cost of around £60000 using a total of 1100 m ofCarboDur laminates.

11.3.2 Strengthening of the Rhine bridge Oberriet,Meiningen, Switzerland


11.3.2.1.1 Introduction

The three span bridge was built in 1963 and crosses the border over theRiver Rhine between Austria and Switzerland, linking the towns ofMeiningen (Vorarlberg) to Oberriet (St Gallen) (Walser and Steiner,1997). The end spans are 35.1m in length with a central span of 45m(Fig. 11.5).

A thorough investigation and structural analysis according to currentSIA (Swiss Engineers and Architects) load standards had shown thatbesides normal maintenance, the bridge deck was in need of transversestrengthening. This was due to the fact that in 1963, the bridge deck hadbeen designed for standard 14 tonne truck loads.

293

SOR11 12/19/98, 4:25 PM293


Figure 11.5 General view of Oberriet bridge.


11.3.2.2.1 Strengthening options

Different solutions were available to guarantee the future structural safetyunder today’s traffic loads.

• replacing the entire bridge deck• improving the cross-section characteristics by providing a concrete

overlay• strengthening by bonding additional reinforcement to the existing deck.

Because the existing concrete slab was in good condition and the chloridecontent exceeded the critical values only in the outer 10mm, it was decided,for economical reasons, not to replace the total deck slab. The additionalconcrete thickness necessary to reach full flexural capacity would, however,have caused unacceptable longitudinal stress. Therefore bonding additionalreinforcement to the deck became the preferred solution. Structural ele-ments strengthened with bonded reinforcement, according to general prac-tice, should have a residual total safety of d � R � 1.2 after failure of theadded strengthening. The fact that the required strengthening factor wasabout 2.15 meant that the sectional area of the deck slab still had to beincreased.

SOR11 12/19/98, 4:25 PM294


Figure 11.6 Cross-section of Oberriet bridge structure.

It was possible to meet all the structural requirements by bonding trans-versal strengthening strips and adding 8cm to the slab thickness (Fig. 11.6).Adding new concrete also allowed the removal of the chloride-contaminated concrete layer by hydro demolition. CarboDur laminates80 mm wide and 1.2mm thick were chosen for strengthening. A total of 160strips, 4 m long, bonded at 75cm intervals were used.

Thanks to this concept, the bridge today is as good as new and fully meetstoday’s stringent safety standards.

11.3.2.2.2 Structural analysis and design

The stress results for the strengthened cross-section are represented in Fig.11.7. The zones with negative bending moments were strengthened withnormal steel reinforcement embedded in the concrete overlay. The stresstransference between the old and new concrete was assured by shear con-nectors. For the dead loading, before bonding of the CarboDur laminates,stress results indicated an almost zero stress situation in the middle of thedeck slab, which leads to the assumption that the existing strain is zero.

Assuming the following values for the calculations:

• Steel II: yield stress, fsy � 350 Nmm�2

• Steel S 500: fsy � 460 Nmm�2

• Concrete 1963: compressive stress, fc � 32.5 Nmm�2

• Concrete 1996: B45/35: fc � 23 Nmm�2

• CarboDur laminate: fLu � 2000Nmm�2, ftk � 3000 Nmm�2.

The results in the middle of the deck slab were as follows:

• Ultimate bending moment before strengthening, MRO � 75KN mm�1

• Ultimate bending moment with added concrete, MR1 � 106KN mm�1

SOR11 12/19/98, 4:25 PM295


Md–167 kNm/m'

133 kNm/m'

15 kNm/m'

Limits MdM permanentd

M /y before strengtheningR RM /y after adding concreteR RM /y final new cross sectionR R

–43 kNm/m'

–50 kNm/m'

65 kNm/m'

88 kNm/m'

Figure 11.7 Transversal bending moments in the deck slab.

Figure 11.8 Determination of the ultimate bending moment.

Cross-section Strains Stresses Forces

Increase concrete areaB45/35

A �=1005 mm

A =772 mm

b=750 mm

A =96 mm

148

d�=

40m

md=

303m

m

h=33

0mm

s

s

ε

x=46

mm

=–1.37‰

ε �=–0.185‰

ε =7.62‰

∆ε =8.42‰

c

s

s

M =136.1 KNmR

σ =–20.7N/mm

σ �=–38.8N/mm

σ =350N/mm

σ =1982N/mm

c

s

LLL

s• • • • •

• • • • •D =421.5KN

D �=39.0KN

Z =270.2KN

Z =190.3KN

c

s

s

L

250

2

2

2 284

46

80

2

2

2 2

• Ultimate bending moment with additional CarboDur strengthening,MR2 � 181.4KN mm�1.

The results for MR2 are represented in Fig. 11.8. The planes of elongationwere determined on the basis of mean elongation values, whereas tensilestresses of steel and CarboDur laminates correspond to maximum elonga-tion for the formulation of the conditions of equilibrium. The followingcoefficients were taken as a ratio of mean to maximum elongation: CFRPstrip KL � 0.7/steel Ks � 0.9.

Figure 11.8 also shows that failure of the CarboDur laminate strip occursduring the yielding of steel but before failure of the concrete. The total

SOR11 12/19/98, 4:25 PM296


strengthening factor of 2.4 is the result of increased concrete thickness loadfactor of 1.4 � CarboDur laminate safety factor of 1.7. Under maximumservice load of mser � 88.9KN mm�1, the requirement that total safetyd � R � 1.2 after failure of the CarboDur laminates is met. The fact thatthe CarboDur laminates do not have any plastic deformation capability isaccounted for by the choice of the value fLu � 2/3 ftk.

11.3.2.2.3 Application of the CarboDur strengthening system

The underside of the bridge deck was prepared by blast cleaning to exposethe coarse and fine aggregate and to give the proper roughness of 0.5–1.0 mmpeak trough amplitude for bonding. To assure a good bond of the epoxyadhesive to the concrete immediately prior to the bonding operations, theconcrete surface was cleaned by vacuum cleaner to remove all dust.

The pretreated substrate was uneven and full of cavities due to debriswithin the original cast concrete. Substrate reprofiling work was carried outwith a compatible epoxy mortar along the proposed bond line, executedon the day preceding the actual bonding operations. Immediately after thefinal cleaning of the substrate, the adhesive was applied by trowel onto theconcrete (Fig. 11.9).

Owing to the lightweight nature of the CarboDur laminates only twooperatives were required to carry the 4.0 m long laminates to the area ofapplication. The two operatives then applied them onto the underside of

Figure 11.9 Application of the Sikadur epoxy adhesive.

SOR11 12/19/98, 4:25 PM297


Figure 11.10 Installation of CarboDur laminate.

Figure 11.11 Photographic cross-section of core taken from bondedlaminate showing adhesive interface and omission of voids, Rhinebridge, Oberriet.

SOR11 12/19/98, 4:25 PM298


Figure 11.12 Determining the concrete surface preparation by a ‘pull-off’ test.

the bridge deck (Fig. 11.10). The CarboDur laminates were then carefullypressed on by means of a hard rubber roller. This method of pressing on byroller has successfully been tested on concrete beams at EMPA (Fig. 11.11).Entrapped air in the adhesive interface was checked by means of infraredthermography.

11.3.2.2.4 Quality assurance

After preparation of the substrate, the surface was inspected visually forweak areas, cracks and inclusions in the concrete such as wood. Tensilebond strengths were used to assess the adequacy of the surface preparationby means of pull-off tests with glued-on steel disks (Fig. 11.12). Preliminaryinvestigations had already shown that the concrete of the deck of theOberriet bridge was of excellent quality. Readings for the tensile bondstrength were in the range of 3.3–3.7 Nmm�2.

The evenness of the concrete surface was checked with a metal straightedge (Fig. 11.13). The maximum allowable deviation of 5 mm over a lengthof 2 m and 1mm over a length of 0.3m was exceeded for 10% of the surface.Such areas were reprofiled by levelling with a compatible Sika epoxymortar to the prescribed admissible tolerances.

To assess site mixing during the bonding operations, prisms of the epoxyadhesive, two per day, 12 in total, were prepared for laboratory testing to

SOR11 12/19/98, 4:25 PM299


Figure 11.13 Checking the evenness of the reprofiled surface.

Figure 11.14 Completed strengthening work on Rhine bridge, Oberriet.

SOR11 12/19/98, 4:25 PM300


measure the compressive strength and flexural modulus. A final inspectionof the applied plates incorporated a visual inspection and a check for hollowspots by tapping the surface with a small hammer (Fig. 11.14).

11.3.3 Strengthening of masonry walls in an officebuilding, Zurich, Switzerland


11.3.3.1.1 Introduction

Two existing six storey apartment houses built in the 1930s were convertedinto a large office building. Consequently, a complete redesign of the struc-tural load bearing system was necessary. Furthermore, many items of thepresent building codes had to be taken into consideration because theydiffered considerably from those at the time of the original construction, inparticular with respect to the earthquake and wind load standards.

Amongst many other alterations, old wooden floors and all of the innerload bearing walls and one entire façade had to be removed and replacedby reinforced concrete slabs and columns. Only parts of the interiorunreinforced masonry (URM) fire wall remained in place. These alterationschanged both the stiffness and the load bearing capacity of the wholestructure.

In the longitudinal direction, two new concrete walls were calculated toresist the earthquake loads. But for the critical transversal direction, onlytwo internal concrete load bearing walls of the staircase and parts of theURM fire wall were available to transmit the horizontal loads down into thefoundation. The interior URM fire wall had, therefore, to be strengthenedconsiderably.


11.3.3.2.1 Strengthening options

Three options were considered:

1 demolishing and reconstructing a new fire wall,2 strengthening the existing wall by applying a reinforced shotcrete skin,3 strengthening the wall using the CarboDur carbon fibre strengthening

system.

The carbon fibre option was chosen based on the following advantages:

• creates minimum interference with other construction work

SOR11 12/19/98, 4:25 PM301


• no dimensional changes in wall thickness• cost effective solution to resist earthquake loads• maintenance free system• no special tools or heavy equipment required on site• short duration time on site resulting in a reduction in programme time.

11.3.3.2.2 Strengthening details

The CarboDur strengthening system was utilised on one side of the wall forthree storeys (Fig. 11.15) using 100mm wide strips, 1.2mm in thicknesslaid diagonally across the wall (Fig. 11.16). The practical advantage of usingthe CarboDur laminates was lightness of the material. One long length oflaminate was used, therefore, eliminating the need for lap joints. In additionthe crossover detail was very simple.

The existing render was removed from the wall and the surface was gritblasted in the areas to be bonded to achieve an open textured profile. Localprotuberances were removed mechanically. Prior to the application of theadhesive, the surfaces were vacuum cleaned to remove dust.

The CarboDur laminates were anchored in the adjacent new reinforcedconcrete column. In order to achieve optimal adhesion between the carbonfibre laminate and the grouting mortar, the anchorage zone of the laminateswas slightly curved and provided with a special bonding bridge. In additionto this, steel ties were placed across the laminates and fixed into the con-crete with an epoxy resin. All anchorage zones were then grouted with acompatible epoxy mortar.

To ensure the highest quality of work, specialist contractors were usedtogether with the following on site quality assurance programme:

• continuous visual over-all inspection by a supervising engineer,• adhesion tests on the prepared surfaces,• dewpoint control of the substrates prior to bonding and grouting,• sampling of all epoxy batches, as used and mixed on the site, to measure

compressive strength and flexural modulus,• recording of all delivery documents, including production numbers and

expiry dates.

11.3.3.2.3 Conclusion

With this strengthening, the lateral resistance and the ductility of the inte-rior URM fire wall could be increased many times over at reasonable costs.It took no more than four days to carry out all the strengthening work. Thiswas the first ever project where carbon fibre laminates had been used tostrengthen a masonry wall (Fig. 11.17).

SOR11 12/19/98, 4:25 PM302


Figure 11.15 Cross-section of the Zurich office building.

SOR11 12/19/98, 4:25 PM303


Figure 11.16 Application of the CarboDur plates to the masonry wall.

304

Figure 11.17 URM fire wall strengthened with CarboDur carbon fibrestrengthening system.

SOR11 12/19/98, 4:25 PM304


11.3.4 Strengthening of historic wooden bridge,Switzerland


11.3.4.1.1 Historical background

In 1807 a covered wooden bridge near Sins in Switzerland was designedand constructed to allow horse drawn vehicles to cross over the river(Fig. 11.18). The original supporting structure design consisted of archesstrengthened by suspended truss members. On the western side of thebridge this construction can still be seen and is currently in good condition.

During the Civil War in 1847 because the bridge was identified as astrategic crossing point, the eastern side was partially destroyed. In 1852,this section of the bridge was rebuilt with a modified supporting structuremade up from a combination of suspended truss members with interlockingtensioning transoms. The present permitted vehicle load carrying capacityof the bridge is 20 tonnes.

During its life the bridge has been rehabilitated in different ways. Themost recent investigation which consisted of a load test identified thatthe wooden pavement and several cross beams were incapable of carryingthe current vehicle loading requirements.

Figure 11.18 General view of wooden bridge, Sins, Switzerland.

SOR11 12/19/98, 4:26 PM305



11.3.4.2.1 Strengthening details

In 1992, strengthening work commenced on the bridge and the originalwooden pavement was replaced with 20cm thick transversely prestressedbonded wooden planks. The most highly loaded cross beams were streng-thened using carbon fibre plates bonded to the external surfaces. Each crossbeam was constructed from two oak beams placed upon each other sepa-rated by wooden blocks to increase the depth (Fig. 11.19). The dimensionsof the upper beam were 300 mm by 300mm and of the lower beam 370mmdeep by 300mm wide.

To achieve the required bonding surface, the beams had to be prepared.The most suitable method of preparation for this project was achieved byusing a portable planing system (Fig. 11.20). The cross beams were streng-thened with 1.0mm thick plates at widths of 250mm at the upper leveland 200 mm at the lower level (Fig. 11.21). Once the preparation wascarried out and dust removed, the plates were bonded to the beams(Fig. 11.22).

Figure 11.19 Cross-section of the historic bridge near Sins. Selectedcross beams were strengthened with carbon fibre plates.

SOR11 12/19/98, 4:26 PM306


Figure 11.20 Portable system to plane the surface of the woodencross beams.

Figure 11.21 View of strengthened cross beams.

SOR11 12/19/98, 4:26 PM307


Figure 11.22 A close-up of the bonded plate to the underside of thebeam.

Figure 11.23 The use of bonded gauge studs to monitor long termperformance.

SOR11 12/19/98, 4:26 PM308


The Swiss Federal Laboratories for Material Testing and Research(EMPA) used pulse infrared thermography to assess the in situ suitability ofthe bonding operation. The use of strain gauges and gauge studs to monitorthe long term performance of the strengthening technique is currently beingused on selected cross beams (Fig. 11.23).

11.3.4.2.2 Conclusion

The success of this project has given confidence and practical experience inthis method for poststrengthening timber structures. For the future preser-vation of historic bridges and similar structures, this poststrengtheningtechnique offers many advantages over traditional strengthening methods.Because carbon fibre plates are thin, strong and flexible, they can be de-signed and installed to provide a cost effective solution which does notdetract visually from the original design of the structure.

11.3.5 Strengthening subways, Tyne and Wear, UK


Following an assessment of the structure by South Tyneside MBC it wasnecessary to upgrade the flexural loading capacity of subways to accommo-date 40 tonne vehicle loadings.

Figure 11.24 Applying CarboDur laminate to underside of underpass,Tyne and Wear.

SOR11 12/19/98, 4:26 PM309


Figure 11.25 View of completed bonding operation, subways, Tyneand Wear.


To increase the flexural capacity, 100mm wide, 1.2mm thick carbon fibrelaminates were bonded to the roof of the subway. The CarboDur laminatesand adjacent concrete roof were overlaid with 15.0mm of prebagged poly-mer modified Gunite and coated with a high performance protective con-crete coating. The total length was 130m (Figs. 11.24 and 11.25).

11.3.6 Underpass at Great Missenden,Buckinghamshire, UK


The underpass was understrength to sustain current loading requirementsof 40 tonnes.

11.3.6.2 Solution

Cracks in the soffit of the deck were first injected with an epoxy resin. Minorconcrete repairs were also carried out prior to bonding the 100mm wide,1.2mm thick CarboDur laminates. The total length was 172m (Figs. 11.26and 11.27).

SOR11 12/19/98, 4:26 PM310


Figure 11.26 General view of the underpass, Great Missenden.

Figure 11.27 Completed bonding operation of underpass, GreatMissenden.

SOR11 12/19/98, 4:26 PM311


11.3.7 Strengthening of loggia slabs (balconies),Magdeburg, Olvenstedt, Germany


All the slabs in a multistorey block of flats were sagging due to fatigue andinsufficient reinforcement. This reduced the safe use of the balconies. Toenable levelling of the upper surface and to compensate for the new deadload and existing live loads, the balconies needed strengthening.


The underside of the loggia slabs was strengthened with three 50mmwide, 1.2mm thick carbon fibre laminates. The total length was 1140m(Figs. 11.28 and 11.29).

Figure 11.28 View showing block of flats, Magdeburg.

SOR11 12/19/98, 4:26 PM312


Figure 11.29 Underside of deck showing completed strengthening ofbalconies, Magdeburg.

11.3.8 Strengthening of floors in an old apartment house,Budapest, Hungary


The owner changed the use of the building from an apartment house to anoffice building. The existing floor slab construction was not sufficient for thenew loading requirements and had to be strengthened.


The floor slab soffit was strengthened with 50mm wide, 1.2mm thickCarboDur laminates. The total length was 170 m (Figs. 11.30 and 11.31).

SOR11 12/19/98, 4:26 PM313


Figure 11.30 Apartment house, Budapest.

Figure 11.31 Application of CarboDur laminates.

SOR11 12/19/98, 4:26 PM314


11.3.9 Strengthening of floor slabs in children’s hospital,Brno Hospital, Czech Republic


The hospital had planned to install a new tomograph and it was found thatthe existing reinforcement was not sufficient for this additional load. Theceiling needed a poststrengthening system to increase the load capacity.


CarboDur laminates 50 mm wide, 1.2 mm thick were applied crosswise tothe underside of the floor. The total length was 30m (Figs. 11.32 and 11.33).

Figure 11.32 Application of CarboDur laminates.

SOR11 12/19/98, 4:26 PM315


Figure 11.33 Completed installation.

11.3.10 Floor strengthening of town hall, Auckland, NewZealand


During the course of the construction work on the town hall it was foundthat the two mezzanine floors in the main entrance area did not havesufficient reinforcement to comply with current codes. The engineersneeded to design a poststrengthening system that could be installed toincrease the live load capacity of the existing floors (Fig. 11.34).


The floor slab soffits were strengthened with 50mm, 1.2 mm thickCarboDur laminates installed at 600mm intervals. The total length was200m (Fig. 11.35).

SOR11 12/19/98, 4:26 PM316


Figure 11.35 View showing plates bonded to structural floor slabsubstrate.

317

Figure 11.34 General view of town hall, Auckland.

SOR11 12/19/98, 4:27 PM317


11.3.11 Strengthening of apartment balconies due to highdeflections: multistorey flats in Loano, Genova,Italy


The cantilever balconies on a multistorey block of flats had an end deflec-tion in excess of 12mm, together with associated cracks in the concrete. Thebalcony slabs had been underdesigned and required strengthening.


Poststrengthening of the balconies was carried out using four plates, 50mmwide, 1.2mm thick and 1.2 m in length on the upper surface of the slab andbeam. A load test was carried out and the deflection was reduced to 2.0mm(Fig. 11.36).

Figure 11.36 Load testing of balcony at Loano.

SOR11 12/19/98, 4:27 PM318


11.3.12 Strengthening of a concrete waffle slab, Stuttgart,Germany


To upgrade the college complex another floor was to be added to thebuilding. To bear the additional load, the original roof slab had to bestrengthened to comply with the new floor loadings.


The slab was strengthened with 500 m of 80mm wide, 1.2 mm thick lami-nates and 250 m of 50mm wide, 1.2 m thick laminates. The thin section ofthe plates allowed the simple detailing of crossovers at the intersections ofthe beams (Figs. 11.37 and 11.38).

Figure 11.37 Offering CarboDur laminates up to the roof beams,Stuttgart.

SOR11 12/19/98, 4:27 PM319


Figure 11.38 The completed project with simple crossover detailsshown, Stuttgart.

11.3.13 Strengthening of longitudinal concrete bridgebeams, Niederwartha near Dresden, Germany


The whole concrete road bridge had to be repaired due to fatigue andenvironmental influences. An assessment of the bridge found that thelongitudinal beams required strengthening to increase the live load capacity(Fig. 11.39).


The beams were strengthened on the underside with a total of 250m, 80mmwide, 1.2mm thick CarboDur laminates and 250m of 50 mm wide, 1.2mmthick CarboDur laminates (Fig. 11.40).

SOR11 12/19/98, 4:27 PM320


Figure 11.40 View of post-strengthened beam.

321

Figure 11.39 General view of repair work, concrete bridge,Niederwarthar near Dresden.

SOR11 12/19/98, 4:27 PM321


11.3.14 Horgen ferry bridge, Switzerland


A deficit of lateral reinforcement in the top deck of this reinforced concretebridge necessitated a strengthening programme. The lack of reinforcementand resultant moment diagram is shown in Fig. 11.41. The shaded area ofthe diagram shows the reduced moment 4.0m from the edge of the bridgeparapet.


To overcome this loading problem, strengthening of the deck transverselywas carried out in 1997 from the top surface using over 700m of CarboDurlaminates (Fig. 11.42). The thickness of the CFRP plates resulted in mini-mum alteration to the structure.

Figure 11.41 Moment diagram for Horgen ferry bridge: lateralmoment, solid line; movement of resistance of steel reinforcement,dashed line; moment deficit, shaded area.

SOR11 12/19/98, 4:27 PM322


11.3.15 Devonshire Place bridge, Skipton, UK


Mouchel was appointed by North Yorkshire County Council to repairDevonshire Place bridge in Skipton, North Yorkshire. The bridge has aprecast prestressed concrete hollow section edge beam. A number of thetendons in the edge beam were damaged during an inspection, weakeningthe edge of the bridge in flexure.


Using knowledge gained from the extensive work done on ROBUST,a single sheet of CarboDur laminate plate was bonded to the undersideof the bridge to replace the lost flexural capacity. The traditional approachof bonding steel plates was clearly not suitable for this bridge due toaccess restrictions. In addition, the existing concrete was not thick enoughto support anchor bolts which would have been required if steel plateswere used. The CarboDur laminate was bonded to the bridge with Sikadur30 adhesive. The plate bonding required no bolts or scaffolding andthe bridge remained open during the process which was completed withinone day.

2400 mm 3200 mm

2000 mm

6000 mm

1300

mm

CarboDur CFRP Strip

Iv Iv

Figure 11.42 Lateral poststrengthening of the deck with a total of700m of CarboDur CFRP strips.

SOR11 12/19/98, 4:27 PM323


11.3.16 Nestlé chocolate factory, Tutbury, UK


Mouchel was commissioned by Nestlé to inspect and access the capacity ofa number of main beams supporting a factory floor in Tutbury to cater fora 30% increase in floor loading due to the installation of new plant andprocessing equipment. The brief also stipulated that there was to be mini-mal disruption to the factory operations.


Mouchel designed the strengthening scheme utilising the CarboDur lami-nate plate. In total 11 beams were required to have their flexural capacityincreased by 30% to cater for the extra loading. This was achieved bybonding one strip of CarboDur laminate plate, 120mm wide by 1.2 mmthick, centrally along the soffit of each of the beams. Around 100m ofCFRP plate was required for the entire operation. No scaffolding equip-ment was required, due to the lightness of the CarboDur plates, whichmeant minimum disruption to the factory operations and short contractduration. Sikadur 30 adhesive was used to bond the plates to the soffits ofthe beams. The carbon fibre composite bonding operation was completed inless than two days.

11.4 References

Farmer N (1997) ‘Strengthening with CFRP laminates’, Construct Repair II(7) 2–4.Midwinter K N (1997) ‘Plate bonding carbon fibre and steel plates’, Construct

Repair II(7) 5–9.Meier U (1995) ‘Strengthening of structures using carbon fibre/epoxy composites’,

Construct Building Mater 9(6) 341–351.Schwegler G (1994) ‘Masonry construction strengthened with fibre composites

in seismically endangered zones’, 10th European Conference on EarthquakeEngineering, Vienna.

Shaw M (1997) ‘Structural strengthening with external plate bonding’, InternationalConference on Building Envelope Systems & Technology, Bath, UK, April 1997.

Walser R and Steiner W (1997) ‘Strengthening of the Rhine bridge Oberriet-Meiningen’, Proc. Recent Advances in Bridge Engineering – Advanced rehabilita-tion, durable materials, non-destructive evaluation and management, ed. U Meierand R Betti, July 1997, pp 127–133.

SOR11 12/19/98, 4:27 PM324

Index 325

325

Index

3M adhesive, 104absorption of water into adhesives and

adherends, 161adherends, 170

surface preparations for, 72–6adhesives, 12, 57, 61, 62adhesive

application, 76–7, 261behaviour (under load) of, 164bond strength, 15bonding characteristics, 47bonding to CFRP plates – site

requirements, 272–6concrete interfaces, 174curing, 77, 264durability, 80effects of environment on joints and

interfaces, 173heat distortion temperature, 264joints, effects of surface treatment on,

173mechanical properties, 264mixing, 76–7, 264moisture resistance, 264placing, 264properties, 61–7quality control, 261selection, 160, 259storage time, 264structural, 12tests, 67, 263–6thickness effects of, 106usable life of, 61, 264

adhesion, 68promoters, 61primers, 61

aliphatic polyamines, 60aramid fibre, 48, 50, 55

composite (bi-directional), 55composite (unidirectional fibres), 55reinforced plastics, 55, 74

anchorage of composite plate, 101–6applications of FRP strengthening, 36ASTM (1983) and (1984), 68Auckland New Zealand floor strengthening,

316–17autoclave moulding technique, 50

bi-linear kinematic hardening (BKIN) –FEA option, 225–6

Boeing wedge cleavage test, 25bonded joints (experimental

considerations), 164bonding of

anchored plates, 280–2multi-plates, 281prestressing CFRP plates, 282unanchored plates, 276–80undrilled plates, 281

bonding operation, 75bond line thickness control, 77Brno hospital Czech Republic floor slab

strengthening, 315–16BS EN 29142 (1993), 163Budapest strengthening of floor, 313–14

CarboDur strengthening system,application of, 297

carbon fibre, 1, 9, 48, 49, 54, 80CFRP, 24, 54–5

Unidirectional, 53, 171Toray T700, 54

composite behaviour under load, 64composite material, 47, 52, 258–9

behaviour under load, 164deterioration of, 171environmental effect on, 166plate end geometry, 25unidirectional, 54

cost comparison of steel and compositeplate bonding, 22

coupling agent, 61creep

of FRP composites, 9, 56, 194–6of structural adhesives, 184–7of steel plated beams, 192–4testing of CFRP material, 197testing of CFRP plated beam, 197–200

de-icing salts, 3

SORINDEX 12/19/98, 3:43 PM325

326 Index

degradation, processes involved in joint,162

Department of Transport, 17diffusion of water into adhesive and

adherends, 161Dresden concrete bridge beam

strengthening, 320–1Drucker–Prager FEA option, 225–6durability, 8, 9, 256

of concrete, 156main factors affecting, 159

dynamic fatigue, 200

economics of CFRP plated beams, 286environmental conditions, effects on

adhesive joints, 157, 163assessing degradation on joints, 163durability of adhesive joints, 156, 160moisture on adhesive joints, 161

end anchorage, 255end anchorage peel stresses, 246–50environmental effects on

fatigue behaviour, 202small scale specimens, 178

EMPA, 18, 24, 31epoxy resins, 1, 12, 46, 48, 54, 58, 59epoxy resin additives, 60epoxy resin rubber toughened, 29EUROCOMP, 74

failure mode of plated RC beam, 243–6fatigue, discussion of plated beam tests

under, 216–17fatigue

of adhesives, 203–6of aramid composites, 55of FRP material, 206–8of FRP plated short span beams, 211–13of FRP plated long span, 213–15, 255of post tensioned prestressed beams, 200of R.C. concrete bridge decks, 201of steel plated beams, 208–10, 255

fatigue, Dynamic, 200fatigue of E-glass composites, 55finite element analysis of

18 metre beams, 231–3effect of adhesive material on beams, 231effect of mesh density on beams, 229–31plated beam, 2D, 225–6plated beam, 3D, 227–9prestressed concrete, 223RC, 223results of prestressed plated beams, 237–9results of unstressed plated beam, 234–7

finite element analysis, crack approach forconcrete

discrete, 224smeared, 224

fire comparison, steel and CFRP platedbeams, 23

fire resistance, 6, 80, 256of vinylester, 179–81

freeze thaw resistance, 6

FRP, fatigue of, 206–8adherends, 170deterioration of mechanical property,

170surface environment on properties –

chemical, 170; mechanical, 170;physical, 170

FRP plateanchorage, 18separation, 18

Fujimi bridge culvert, Japan, 38

galvanic corrosion, 80glass fibre, 1, 48, 55

A, 48, 49, 54E, 48, 49, 54, 55environmental attack on, 172R, 48, 49, 54reinforcement, 9S, 49, 54

glass transition temperature, 58, 63, 163,164, 168, 171

Great Missenden underpass strengthening,310–11

grit blasting, 6

heat distortion temperature (HDT), 67, 265heat distortion tests, 265Hata bridge Japan, 38Horgen ferry bridge strengthening, 322–3

Ibach bridge, 36inspection and maintenance of strengthened

structures, 266interface – FRP/adhesive joint, 176isophthalic polyesters, 158, 172

joints, adhesivedegradation of, 162effect of environment on, 173environmental and service conditions, 157factors affecting joint durability, 158surface treatment of, 173

joints, adhesive/concrete interfaces, 174joint performance

fire, 179freeze/thaw, 179

Kevlar 29, 50Kevlar 49, 50, 55Kings College Hospital roof strengthening,

37, 290–3

lap joint test, 262limit state for strengthening beams, 254limit state partial safety factors, 254Loano Genova apartment balconies

strengthening, 318

Magdeburg Olvenstedt strengthening ofbalconies, 312–13

maintenanceof CFRP plated beams, 286

SORINDEX 12/19/98, 3:43 PM326

Index 327

of strengthened structure, 266masonry walls, strengthening of, 301–4methyl ethyl ketone (MEK), 73mixing requirements for adhesives, 264moisture resistance of adhesives, 264

Oberriet Meiningen bridge, Switzerland,293–301

structural analysis and design of, 295orthophthalic polyester, 158, 172

partial safety factors, limit state, 254peel

at end anchorage, 246–50at wide shear/yield crack, 250–2ply, 9, 52, 73, 260stress, 47test, modified, 16

plate end geometry, 100plate stiffness, effect of increasing, 106polyacrylonitrile PAN fibre, 49polyester resin, 12, 48polyester, unsaturated, 13prepreg, 50prepreg tapes, 99prestressing of CFRP plates, 31, 282–6

externally bonded plates, 136laboratory techniques for, 138–40

prestressed concrete beams, testing of, 113prestressed plates, 135

segmental construction using, 135strengthening of prestressed concrete

beams by, 135properties

of Sika CarboDur material, 289of Sikadur 30, 289

pultrusion, 51–2CFRP plates, 54

pull-off tests PrEN 1542 (1996), 174quality control of adhesives, 261

Quinton bridge, 17

reinforced concrete beamsplated, 85–91unplated, 84

RILEM 15

shearconnection in steel/concrete construction,

16effects in R.C. theory, 252lap joints, 105span/beam depth ratio, 88, 91, 93, 94, 97,

98, 100, 105, 108, 252strengthening, 34stress between plate and concrete,

longitudinal, 252test slant, 15, 16, 165

Sikadur 31 PBA, propertiescreep, 187environmental, 166, 168interfacial failure, 169

material characteristics, 169mechanical, 63moisture absorption, 168thermal cycling, 168toughness of joints, 169

Silane surface primer, 16Sin wooden bridge, 37

strengthening of, 305–9Skipton bridge strengthening, 323steel plates external strengthening, 17steel plate bonding, site technique for, 270–1strengthening of structures

advantages, 4marketing, 2techniques, 3

stress rupture, 56structural adhesives

bond tests, 15requirements for, 14type of, 12

Stuttgart waffle slab strengthening, 319–20surface preparation of adherends, 68, 70, 76,

260

thermal expansion coefficient, 56, 171thermomechanical performance, 48thermosetting resin, 48thick adherend shear test (TAST), 67time dependent characteristics of

adhesive polymer, 184–7composite plated beams, 196concrete, 184FRP composites, 194–6steel, 184steel plated beams, 192–4

Tokando Highway bridge Japan, 38Tougheners, 61Transport Research Laboratory, 17, 18, 193,

201transverse tensile stress, 48Tutbury UK, beam strengthening, 324Tyne and Wear UK, subway strengthening,

309–10

urethane adhesive, 16urethane methacrylate, 48

vacuum bag moulding, 50, 53vandalism, 256vinylester, 1, 46, 48

in fire, 179–81

wedge cleavage test, 262–3

Yoshino Route Tunnel, Japan, 38

zirconia glass, 9Zurich, ground floor strengthening of rail

terminal, 37strengthening of masonry wall,

Switzerland, 301

SORINDEX 12/19/98, 3:43 PM327

Strengthening of Reinforced Concrete Structures: Using Externally-Bonded FRP Composites in...

Documents

Transcript of Strengthening of Reinforced Concrete Structures: Using Externally-Bonded FRP Composites in...